This is an automated email from the ASF dual-hosted git repository.

yiguolei pushed a commit to branch branch-4.1
in repository https://gitbox.apache.org/repos/asf/doris.git


The following commit(s) were added to refs/heads/branch-4.1 by this push:
     new 1512485c45c branch-4.1:[fix](rowid-fetch) fix TopN lazy 
materialization crash with duplicateprojected columns (#64372) (#67563)
1512485c45c is described below

commit 1512485c45c723db2579ba3959809616491f28a1
Author: daidai <[email protected]>
AuthorDate: Thu Oct 8 19:10:22 2026 +0800

    branch-4.1:[fix](rowid-fetch) fix TopN lazy materialization crash with 
duplicateprojected columns (#64372) (#67563)
    
    ### What problem does this PR solve?
    Problem Summary:
    bp #64372
    
    ### Release note
    
    None
    
    ### Check List (For Author)
    
    - Test <!-- At least one of them must be included. -->
        - [ ] Regression test
        - [ ] Unit Test
        - [ ] Manual test (add detailed scripts or steps below)
        - [ ] No need to test or manual test. Explain why:
    - [ ] This is a refactor/code format and no logic has been changed.
            - [ ] Previous test can cover this change.
            - [ ] No code files have been changed.
            - [ ] Other reason <!-- Add your reason?  -->
    
    - Behavior changed:
        - [ ] No.
        - [ ] Yes. <!-- Explain the behavior change -->
    
    - Does this need documentation?
        - [ ] No.
    - [ ] Yes. <!-- Add document PR link here. eg:
    https://github.com/apache/doris-website/pull/1214 -->
    
    ### Check List (For Reviewer who merge this PR)
    
    - [ ] Confirm the release note
    - [ ] Confirm test cases
    - [ ] Confirm document
    - [ ] Add branch pick label <!-- Add branch pick label that this PR
    should merge into -->
    
    Co-authored-by: York Cao <[email protected]>
    Co-authored-by: caokaihua1 <[email protected]>
---
 be/src/exec/rowid_fetcher.cpp                      |  267 +++--
 be/src/exec/rowid_fetcher.h                        |   15 +-
 be/test/exec/rowid_fetcher_test.cpp                |  167 ++++
 .../hive/test_hive_topn_lazy_mat.out               | 1047 ++++++++++++++++++++
 .../tvf/test_tvf_topn_lazy_mat.out                 |   72 ++
 .../hive/test_hive_topn_lazy_mat.groovy            |   50 +-
 .../tvf/test_tvf_topn_lazy_mat.groovy              |    7 +
 7 files changed, 1552 insertions(+), 73 deletions(-)

diff --git a/be/src/exec/rowid_fetcher.cpp b/be/src/exec/rowid_fetcher.cpp
index 92f8de5c2e6..0219ac0b778 100644
--- a/be/src/exec/rowid_fetcher.cpp
+++ b/be/src/exec/rowid_fetcher.cpp
@@ -32,6 +32,7 @@
 #include <cstdint>
 #include <limits>
 #include <memory>
+#include <mutex>
 #include <ostream>
 #include <string>
 #include <unordered_map>
@@ -907,16 +908,22 @@ Status RowIdStorageReader::read_lance_rows_by_row_ids(
 
 Status RowIdStorageReader::read_external_row_from_file_mapping(
         size_t idx, const std::multimap<uint64_t, size_t>& row_ids,
-        const std::shared_ptr<FileMapping>& file_mapping, const 
std::vector<SlotDescriptor>& slots,
-        const TUniqueId& query_id, const std::shared_ptr<RuntimeState>& 
runtime_state,
-        std::vector<Block>& scan_blocks, std::vector<std::pair<size_t, 
size_t>>& row_id_block_idx,
+        const std::shared_ptr<FileMapping>& file_mapping,
+        const std::vector<SlotDescriptor>& scan_slots, const TUniqueId& 
query_id,
+        const std::shared_ptr<RuntimeState>& runtime_state, 
std::vector<Block>& scan_blocks,
+        std::vector<std::pair<size_t, size_t>>& row_id_block_idx,
         std::vector<RowIdStorageReader::ExternalFetchStatistics>& 
fetch_statistics,
         const TFileScanRangeParams& rpc_scan_params,
         const std::unordered_map<std::string, int>& colname_to_slot_id,
-        TupleDescriptor& tuple_desc) {
+        std::counting_semaphore<>& semaphore, TupleDescriptor& tuple_desc) {
     SCOPED_ATTACH_TASK(ExecEnv::GetInstance()->rowid_storage_reader_tracker());
     signal::set_signal_task_id(query_id);
 
+    // Release the concurrency permit on every exit path (including error 
returns
+    // and exceptions). Completion accounting and status publishing are owned 
by
+    // the caller, so the status is always published before the waiter is 
woken.
+    Defer defer([&] { semaphore.release(); });
+
     std::vector<uint64_t> read_ids;
     // Generate an ordered, deduplicated list with the help of the multimap 
ordering.
     for (const auto& [row_id, result_block_idx] : row_ids) {
@@ -926,7 +933,7 @@ Status 
RowIdStorageReader::read_external_row_from_file_mapping(
         row_id_block_idx[result_block_idx] = std::make_pair(idx, 
read_ids.size() - 1);
     }
 
-    scan_blocks[idx] = Block(slots, read_ids.size());
+    scan_blocks[idx] = Block(scan_slots, read_ids.size());
 
     auto& external_info = file_mapping->get_external_file_info();
     auto scan_range_desc = 
build_external_fetch_range(external_info.scan_range_desc);
@@ -936,9 +943,10 @@ Status 
RowIdStorageReader::read_external_row_from_file_mapping(
     const auto format_type = scan_range_desc.__isset.format_type ? 
scan_range_desc.format_type
                                                                  : 
rpc_scan_params.format_type;
     if (format_type == TFileFormatType::FORMAT_LANCE) {
-        RETURN_IF_ERROR(read_lance_rows_by_row_ids(
-                scan_range_desc, read_ids, slots, runtime_state.get(), 
sub_runtime_profile.get(),
-                rpc_scan_params, &scan_blocks[idx], &fetch_statistics[idx]));
+        RETURN_IF_ERROR(read_lance_rows_by_row_ids(scan_range_desc, read_ids, 
scan_slots,
+                                                   runtime_state.get(), 
sub_runtime_profile.get(),
+                                                   rpc_scan_params, 
&scan_blocks[idx],
+                                                   &fetch_statistics[idx]));
     } else {
         // Parquet/ORC IDs remain signed physical file positions; Lance keeps 
native uint64 IDs.
         std::list<int64_t> legacy_read_ids;
@@ -970,6 +978,23 @@ Status 
RowIdStorageReader::read_external_row_from_file_mapping(
         }
     }
 
+    if (scan_blocks[idx].rows() != read_ids.size()) {
+        return Status::InternalError(
+                "Row id fetch scan row count mismatch, "
+                "query_id={}, path={}, expected_rows={}, actual_rows={}",
+                print_id(query_id), scan_range_desc.path, read_ids.size(), 
scan_blocks[idx].rows());
+    }
+    for (size_t column_id = 0; column_id < scan_blocks[idx].columns(); 
++column_id) {
+        const auto& column = scan_blocks[idx].get_by_position(column_id);
+        if (column.column->size() != read_ids.size()) {
+            return Status::InternalError(
+                    "Row id fetch scan column row count mismatch, "
+                    "query_id={}, path={}, column={}, expected_rows={}, 
actual_rows={}",
+                    print_id(query_id), scan_range_desc.path, column.name, 
read_ids.size(),
+                    column.column->size());
+        }
+    }
+
     auto file_read_bytes_counter =
             
sub_runtime_profile->get_counter(FileScannerV2::FileReadBytesProfile);
 
@@ -988,6 +1013,92 @@ Status 
RowIdStorageReader::read_external_row_from_file_mapping(
     return Status::OK();
 }
 
+std::string RowIdStorageReader::source_column_key(const SlotDescriptor& slot, 
uint32_t column_idx) {
+    fmt::memory_buffer key;
+    // Length-prefix each component so distinct sequences cannot alias, e.g.
+    // paths ["a", "b"] -> "1:a1:b" while ["a:b"] -> "3:a:b".
+    auto append = [&key](std::string_view component) {
+        fmt::format_to(key, "{}:", component.size());
+        key.append(component.data(), component.data() + component.size());
+    };
+    append(slot.col_name());
+    append(std::to_string(column_idx));
+    append(std::to_string(slot.col_unique_id()));
+    append(std::to_string(slot.column_paths().size()));
+    for (const auto& path : slot.column_paths()) {
+        append(path);
+    }
+    append(std::to_string(slot.all_access_paths().size()));
+    // Encode each optional sub-path's presence bit separately from its element
+    // count so an absent path ("0") never aliases a present-but-empty path
+    // ("1" + size "0").
+    auto append_optional_path = [&append](bool is_set, const 
std::vector<std::string>& items) {
+        append(is_set ? "1" : "0");
+        if (is_set) {
+            append(std::to_string(items.size()));
+            for (const auto& item : items) {
+                append(item);
+            }
+        }
+    };
+    for (const auto& path : slot.all_access_paths()) {
+        append(fmt::format("{}", path.type));
+        append_optional_path(path.__isset.data_access_path, 
path.data_access_path.path);
+        append_optional_path(path.__isset.meta_access_path, 
path.meta_access_path.path);
+    }
+    return fmt::to_string(key);
+}
+
+Status RowIdStorageReader::submit_external_scan_tasks(
+        ScannerScheduler* scheduler, std::counting_semaphore<>& semaphore, 
size_t task_count,
+        const std::function<std::string(size_t)>& make_task_id,
+        const std::function<Status(size_t)>& run_task) {
+    // `completed_count` is a plain counter guarded by `mtx`; the same mutex 
guards
+    // the wait predicate below, so a worker can never notify between the 
waiter's
+    // predicate check and its wait.
+    AtomicStatus scan_status;
+    std::condition_variable cv;
+    std::mutex mtx;
+    size_t completed_count = 0;
+
+    // Only tasks the scheduler actually accepted are waited for. If a 
submission
+    // fails we stop submitting, but still wait for the already-accepted tasks 
so
+    // their workers cannot outlive the locals they capture by reference.
+    size_t submitted_count = 0;
+    for (size_t idx = 0; idx < task_count; ++idx) {
+        semaphore.acquire();
+        auto run_one_task = [&, idx]() -> bool {
+            Status task_status = Status::OK();
+            // Publish the status before the completion signal wakes the 
waiter, on every
+            // path. A scanner that throws would otherwise leave scan_status 
OK while this
+            // Defer still counts the task as finished, and the caller would 
report success
+            // over a half-filled result block.
+            Defer complete([&] {
+                scan_status.update(task_status);
+                std::lock_guard<std::mutex> lock(mtx);
+                ++completed_count;
+                cv.notify_one();
+            });
+            ASSIGN_STATUS_IF_CATCH_EXCEPTION(task_status = run_task(idx), 
task_status);
+            return true;
+        };
+        Status submit_st = scheduler->submit_scan_task(
+                SimplifiedScanTask(run_one_task, nullptr, nullptr), 
make_task_id(idx));
+        if (!submit_st.ok()) {
+            scan_status.update(submit_st);
+            semaphore.release();
+            break;
+        }
+        ++submitted_count;
+    }
+
+    {
+        std::unique_lock<std::mutex> lock(mtx);
+        cv.wait(lock, [&] { return completed_count == submitted_count; });
+    }
+    return scan_status.ok() ? Status::OK() : scan_status.status();
+}
+
 Status RowIdStorageReader::read_batch_external_row(
         const uint64_t workload_group_id, const PRequestBlockDesc& 
request_block_desc,
         std::shared_ptr<IdFileMap> id_file_map, std::vector<SlotDescriptor>& 
slots,
@@ -998,12 +1109,21 @@ Status RowIdStorageReader::read_batch_external_row(
     TupleDescriptor tuple_desc(request_block_desc.desc(), false);
     std::unordered_map<std::string, int> colname_to_slot_id;
     std::shared_ptr<RuntimeState> runtime_state = nullptr;
+    std::vector<SlotDescriptor> scan_slots;
+    std::vector<size_t> result_column_to_scan_column;
+    std::vector<uint32_t> scan_column_idxs;
 
     int max_file_scanners = 0;
     {
         if (result_block.is_empty_column()) [[likely]] {
             result_block = Block(slots, request_block_desc.row_id_size());
         }
+        if (request_block_desc.column_idxs_size() != slots.size()) {
+            return Status::InternalError(
+                    "Row id fetch request has mismatched slots and column 
indexes, "
+                    "query_id={}, slots={}, column_idxs={}",
+                    print_id(query_id), slots.size(), 
request_block_desc.column_idxs_size());
+        }
 
         auto& external_info = first_file_mapping->get_external_file_info();
         int plan_node_id = external_info.plan_node_id;
@@ -1028,16 +1148,30 @@ Status RowIdStorageReader::read_batch_external_row(
             }
         }
 
-        std::vector<uint32_t> scan_column_idxs;
+        std::unordered_map<std::string, size_t> source_column_to_scan_idx;
+
+        result_column_to_scan_column.reserve(slots.size());
+        scan_slots.reserve(slots.size());
         scan_column_idxs.reserve(slots.size());
-        for (int slot_idx = 0; slot_idx < slots.size(); ++slot_idx) {
-            auto& slot = slots[slot_idx];
+        for (auto slot_idx = 0; slot_idx < slots.size(); ++slot_idx) {
+            const auto& slot = slots[slot_idx];
+            const auto column_idx = request_block_desc.column_idxs(slot_idx);
+            const auto key = source_column_key(slot, column_idx);
+            auto [it, inserted] =
+                    source_column_to_scan_idx.emplace(key, 
source_column_to_scan_idx.size());
+            result_column_to_scan_column.emplace_back(it->second);
+            if (inserted) {
+                scan_slots.emplace_back(slot);
+                scan_column_idxs.emplace_back(column_idx);
+            }
+        }
+
+        for (auto& slot : scan_slots) {
             tuple_desc.add_slot(&slot);
-            colname_to_slot_id.emplace(slot.col_name(), slot.id());
-            
scan_column_idxs.emplace_back(request_block_desc.column_idxs(slot_idx));
+            colname_to_slot_id[slot.col_name()] = slot.id();
         }
-        rpc_scan_params = build_external_scan_params(rpc_scan_params, 
first_scan_range_desc, slots,
-                                                     scan_column_idxs);
+        rpc_scan_params = build_external_scan_params(rpc_scan_params, 
first_scan_range_desc,
+                                                     scan_slots, 
scan_column_idxs);
 
         const auto& query_options = id_file_map->get_query_options();
         const auto& query_globals = id_file_map->get_query_globals();
@@ -1131,75 +1265,68 @@ Status RowIdStorageReader::read_batch_external_row(
     int64_t scan_running_time = 0;
     RETURN_IF_ERROR(scope_timer_run(
             [&]() -> Status {
-                // Make sure to insert data into result_block only after all 
scan tasks have been executed.
-                std::atomic<int> producer_count {0};
-                std::condition_variable cv;
-                std::mutex mtx;
-                Status scan_status = Status::OK();
-
                 //semaphore: Limit the number of scan tasks submitted at one 
time
                 std::counting_semaphore semaphore {max_file_scanners};
 
-                size_t idx = 0;
+                std::vector<
+                        std::pair<std::multimap<uint64_t, size_t>, 
std::shared_ptr<FileMapping>>>
+                        scan_info_list;
+                scan_info_list.reserve(scan_rows.size());
                 for (const auto& [_, scan_info] : scan_rows) {
-                    semaphore.acquire();
-                    RETURN_IF_ERROR(remote_scan_sched->submit_scan_task(
-                            SimplifiedScanTask(
-                                    [&, idx, scan_info]() -> bool {
-                                        Defer complete_task {[&]() {
-                                            semaphore.release();
-                                            if (++producer_count == 
scan_rows.size()) {
-                                                std::lock_guard<std::mutex> 
lock(mtx);
-                                                cv.notify_one();
-                                            }
-                                        }};
-                                        const auto& [row_ids, file_mapping] = 
scan_info;
-                                        auto status = 
read_external_row_from_file_mapping(
-                                                idx, row_ids, file_mapping, 
slots, query_id,
-                                                runtime_state, scan_blocks, 
row_id_block_idx,
-                                                fetch_statistics, 
rpc_scan_params,
-                                                colname_to_slot_id, 
tuple_desc);
-                                        if (!status.ok()) {
-                                            std::lock_guard<std::mutex> 
lock(mtx);
-                                            if (scan_status.ok()) {
-                                                scan_status = status;
-                                            }
-                                        }
-                                        // The return value indicates whether 
this one-shot
-                                        // scheduler task has completed, not 
whether the fetch
-                                        // succeeded. The fetch status is 
propagated by scan_status.
-                                        return true;
-                                    },
-                                    nullptr, nullptr),
-                            fmt::format("{}-read_batch_external_row-{}", 
print_id(query_id), idx)));
-                    idx++;
+                    scan_info_list.emplace_back(scan_info);
                 }
 
-                {
-                    std::unique_lock<std::mutex> lock(mtx);
-                    cv.wait(lock, [&] { return producer_count == 
scan_rows.size(); });
-                }
-                RETURN_IF_ERROR(scan_status);
-                return Status::OK();
+                return submit_external_scan_tasks(
+                        remote_scan_sched, semaphore, scan_rows.size(),
+                        [&](size_t idx) {
+                            return 
fmt::format("{}-read_batch_external_row-{}", print_id(query_id),
+                                               idx);
+                        },
+                        [&](size_t idx) -> Status {
+                            const auto& [row_ids, file_mapping] = 
scan_info_list[idx];
+                            return read_external_row_from_file_mapping(
+                                    idx, row_ids, file_mapping, scan_slots, 
query_id, runtime_state,
+                                    scan_blocks, row_id_block_idx, 
fetch_statistics,
+                                    rpc_scan_params, colname_to_slot_id, 
semaphore, tuple_desc);
+                        });
             },
             &scan_running_time));
 
-    // Insert the read data into result_block.
-    for (size_t column_id = 0; column_id < result_block.columns(); 
column_id++) {
+    // Insert the read data into result_block. Use insert_indices_from() 
instead of
+    // scatter_scan_blocks_to_result_block()/insert_from_multi_column(), 
because
+    // scan_blocks may have fewer columns than result_block when duplicate 
physical columns
+    // are deduplicated, and insert_from_multi_column() cannot handle 
ColumnString
+    // cross-type (32/64) copies safely.
+    const size_t result_column_count = result_block.columns();
+    for (size_t column_id = 0; column_id < result_column_count; column_id++) {
         auto dst_col_guard = result_block.mutate_column_scoped(column_id);
         MutableColumnPtr& dst_col = dst_col_guard.mutable_column();
 
-        std::vector<const IColumn*> scan_src_columns;
-        scan_src_columns.reserve(row_id_block_idx.size());
-        std::vector<size_t> scan_positions;
-        scan_positions.reserve(row_id_block_idx.size());
+        bool dst_is_nullable = dst_col->is_nullable();
+        std::vector<ColumnPtr> nullable_src_columns(scan_blocks.size());
+        auto scan_column_id = result_column_to_scan_column[column_id];
         for (const auto& [pos_block, block_idx] : row_id_block_idx) {
-            DCHECK(scan_blocks.size() > pos_block);
-            DCHECK(scan_blocks[pos_block].get_columns().size() > column_id);
-            
scan_src_columns.emplace_back(scan_blocks[pos_block].get_columns()[column_id].get());
-            scan_positions.emplace_back(block_idx);
+            DCHECK_GT(scan_blocks.size(), pos_block);
+            DCHECK_GT(scan_blocks[pos_block].columns(), scan_column_id);
+            const auto& src_column_ptr =
+                    
scan_blocks[pos_block].get_by_position(scan_column_id).column;
+            const auto* src_col = src_column_ptr.get();
+            if (dst_is_nullable && !src_col->is_nullable()) {
+                if (!nullable_src_columns[pos_block]) {
+                    nullable_src_columns[pos_block] = 
make_nullable(src_column_ptr);
+                }
+                src_col = nullable_src_columns[pos_block].get();
+            }
+            if (block_idx >= src_col->size()) {
+                return Status::InternalError(
+                        "Row id fetch source index out of range, query_id={}, 
column={}, "
+                        "source_block={}, source_rows={}, row_index={}",
+                        print_id(query_id), 
result_block.get_by_position(column_id).name, pos_block,
+                        src_col->size(), block_idx);
+            }
+            uint32_t scan_position = cast_set<uint32_t>(block_idx);
+            dst_col->insert_indices_from(*src_col, &scan_position, 
&scan_position + 1);
         }
-        dst_col->insert_from_multi_column(scan_src_columns, scan_positions);
     }
 
     // Statistical runtime profile information.
diff --git a/be/src/exec/rowid_fetcher.h b/be/src/exec/rowid_fetcher.h
index fed8e5ccd52..ca1f8b50690 100644
--- a/be/src/exec/rowid_fetcher.h
+++ b/be/src/exec/rowid_fetcher.h
@@ -22,6 +22,8 @@
 #include <gen_cpp/DataSinks_types.h>
 #include <gen_cpp/internal_service.pb.h>
 
+#include <condition_variable>
+#include <functional>
 #include <map>
 #include <memory>
 #include <semaphore>
@@ -42,6 +44,7 @@ class RuntimeProfile;
 class RuntimeState;
 class TQueryOptions;
 class TupleDescriptor;
+class ScannerScheduler;
 namespace io {
 enum class FileCacheMissPolicy : uint8_t;
 }
@@ -172,13 +175,21 @@ private:
     static Status read_external_row_from_file_mapping(
             size_t idx, const std::multimap<uint64_t, size_t>& row_ids,
             const std::shared_ptr<FileMapping>& file_mapping,
-            const std::vector<SlotDescriptor>& slots, const TUniqueId& 
query_id,
+            const std::vector<SlotDescriptor>& scan_slots, const TUniqueId& 
query_id,
             const std::shared_ptr<RuntimeState>& runtime_state, 
std::vector<Block>& scan_blocks,
             std::vector<std::pair<size_t, size_t>>& row_id_block_idx,
             std::vector<ExternalFetchStatistics>& fetch_statistics,
             const TFileScanRangeParams& rpc_scan_params,
             const std::unordered_map<std::string, int>& colname_to_slot_id,
-            TupleDescriptor& tuple_desc);
+            std::counting_semaphore<>& semaphore, TupleDescriptor& tuple_desc);
+
+    static std::string source_column_key(const SlotDescriptor& slot, uint32_t 
column_idx);
+
+    static Status submit_external_scan_tasks(ScannerScheduler* scheduler,
+                                             std::counting_semaphore<>& 
semaphore,
+                                             size_t task_count,
+                                             const 
std::function<std::string(size_t)>& make_task_id,
+                                             const 
std::function<Status(size_t)>& run_task);
 
     struct ExternalFetchStatistics {
         int64_t init_reader_ms = 0;
diff --git a/be/test/exec/rowid_fetcher_test.cpp 
b/be/test/exec/rowid_fetcher_test.cpp
index 2a092a20f70..3e8fd3e9517 100644
--- a/be/test/exec/rowid_fetcher_test.cpp
+++ b/be/test/exec/rowid_fetcher_test.cpp
@@ -33,13 +33,24 @@
 
 namespace doris {
 
+// source_column_key() decides which projected slots collapse onto a single 
scanned
+// column. Two slots that name the same physical column must produce the same 
key, and
+// any two slots that do not must produce different ones -- a false merge 
drops a column
+// the result block still expects, which is the crash this keying was 
introduced to fix.
 class RowIdStorageReaderTest : public testing::Test {
+public:
+    static std::string key_of(const SlotDescriptor& slot, uint32_t column_idx) 
{
+        return RowIdStorageReader::source_column_key(slot, column_idx);
+    }
+
 protected:
     struct SlotSpec {
         std::string col_name = "c";
         int32_t slot_id = 0;
         PrimitiveType type = TYPE_INT;
         int32_t col_unique_id = 1;
+        std::vector<std::string> column_paths = {};
+        TColumnAccessPaths access_paths = {};
     };
 
     static SlotDescriptor make_slot(const SlotSpec& spec) {
@@ -51,8 +62,27 @@ protected:
                                         .build();
         tdesc.__set_id(spec.slot_id);
         tdesc.__set_col_unique_id(spec.col_unique_id);
+        tdesc.__set_column_paths(spec.column_paths);
+        if (!spec.access_paths.empty()) {
+            tdesc.__set_all_access_paths(spec.access_paths);
+        }
         return SlotDescriptor(tdesc);
     }
+
+    static TColumnAccessPath data_path(const std::vector<std::string>& path) {
+        TColumnAccessPath access_path;
+        access_path.type = TAccessPathType::DATA;
+        TDataAccessPath data;
+        data.__set_path(path);
+        access_path.__set_data_access_path(data);
+        return access_path;
+    }
+
+    static TColumnAccessPath bare_path() {
+        TColumnAccessPath access_path;
+        access_path.type = TAccessPathType::DATA;
+        return access_path;
+    }
 };
 
 TEST_F(RowIdStorageReaderTest, ExternalScannerSelectionRespectsRolloutOption) {
@@ -256,4 +286,141 @@ TEST_F(RowIdStorageReaderTest, 
ExternalFetchPartitionSlotsPreserveHivePositionMa
     }
 }
 
+TEST_F(RowIdStorageReaderTest, SameSourceColumnSharesKey) {
+    // The bug case: one physical column projected twice must dedup onto one 
scan column.
+    const SlotDescriptor first = make_slot({});
+    const SlotDescriptor second = make_slot({});
+    EXPECT_EQ(key_of(first, 3), key_of(second, 3));
+}
+
+TEST_F(RowIdStorageReaderTest, ColumnIndexSeparatesKeys) {
+    const SlotDescriptor slot = make_slot({});
+    EXPECT_NE(key_of(slot, 3), key_of(slot, 4));
+}
+
+TEST_F(RowIdStorageReaderTest, ColumnNameSeparatesKeys) {
+    EXPECT_NE(key_of(make_slot({.col_name = "a"}), 0), 
key_of(make_slot({.col_name = "b"}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, UniqueIdSeparatesKeys) {
+    EXPECT_NE(key_of(make_slot({.col_unique_id = 1}), 0),
+              key_of(make_slot({.col_unique_id = 2}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, NameAndIndexBoundaryIsNotAmbiguous) {
+    // Without length prefixes, "a" + idx 12 and "a1" + idx 2 both flatten to 
"a12".
+    EXPECT_NE(key_of(make_slot({.col_name = "a"}), 12), 
key_of(make_slot({.col_name = "a1"}), 2));
+}
+
+TEST_F(RowIdStorageReaderTest, PathComponentBoundaryIsNotAmbiguous) {
+    // The concatenation hazard the length prefix exists for: ["a", "b"] and 
["a:b"] are
+    // different nested columns but share the naive ':'-joined spelling.
+    EXPECT_NE(key_of(make_slot({.column_paths = {"a", "b"}}), 0),
+              key_of(make_slot({.column_paths = {"a:b"}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, EmptyPathIsNotTheSameAsNoPath) {
+    EXPECT_NE(key_of(make_slot({.column_paths = {}}), 0),
+              key_of(make_slot({.column_paths = {""}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, PathOrderMatters) {
+    EXPECT_NE(key_of(make_slot({.column_paths = {"a", "b"}}), 0),
+              key_of(make_slot({.column_paths = {"b", "a"}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, EqualPathsShareKey) {
+    EXPECT_EQ(key_of(make_slot({.column_paths = {"a", "b"}}), 0),
+              key_of(make_slot({.column_paths = {"a", "b"}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, AccessPathSeparatesKeys) {
+    EXPECT_NE(key_of(make_slot({.access_paths = {data_path({"a"})}}), 0),
+              key_of(make_slot({.access_paths = {data_path({"b"})}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, AbsentAccessPathIsNotAnEmptyOne) {
+    // The presence bit: an unset data_access_path must not collide with one 
that is set
+    // but carries no components.
+    EXPECT_NE(key_of(make_slot({.access_paths = {bare_path()}}), 0),
+              key_of(make_slot({.access_paths = {data_path({})}}), 0));
+}
+
+TEST_F(RowIdStorageReaderTest, AccessPathCountSeparatesKeys) {
+    EXPECT_NE(key_of(make_slot({.access_paths = {data_path({"a"})}}), 0),
+              key_of(make_slot({.access_paths = {data_path({"a"}), 
data_path({"b"})}}), 0));
+}
+
+// Runs every submitted task on the submitting thread. The point of these 
cases is which
+// status reaches the caller, not the threading, and inline execution keeps 
them
+// deterministic.
+class InlineScanScheduler : public ScannerScheduler {
+public:
+    Status start(int, int, int, int) override { return Status::OK(); }
+    void stop() override {}
+    Status submit_scan_task(SimplifiedScanTask scan_task) override {
+        scan_task.scan_func();
+        return Status::OK();
+    }
+    Status submit_scan_task(SimplifiedScanTask scan_task, const std::string&) 
override {
+        scan_task.scan_func();
+        return Status::OK();
+    }
+    void reset_thread_num(int, int, int) override {}
+    int get_queue_size() override { return 0; }
+    int get_active_threads() override { return 0; }
+    std::vector<int> thread_debug_info() override { return {}; }
+    Status schedule_scan_task(std::shared_ptr<ScannerContext>, 
std::shared_ptr<ScanTask>,
+                              std::unique_lock<std::mutex>&) override {
+        return Status::OK();
+    }
+};
+
+// submit_external_scan_tasks() signals completion from a Defer, so a worker 
that leaves
+// without publishing its status would still wake the waiter and the caller 
would report
+// success over a partially filled result block.
+class SubmitExternalScanTasksTest : public RowIdStorageReaderTest {
+protected:
+    static constexpr size_t kTaskCount = 3;
+
+    static Status run_tasks(const std::function<Status(size_t)>& run_task) {
+        InlineScanScheduler scheduler;
+        std::counting_semaphore<> semaphore {kTaskCount};
+        return RowIdStorageReader::submit_external_scan_tasks(
+                &scheduler, semaphore, kTaskCount,
+                [](size_t idx) { return fmt::format("task-{}", idx); }, 
run_task);
+    }
+};
+
+TEST_F(SubmitExternalScanTasksTest, AllTasksSucceedingReturnsOk) {
+    size_t ran = 0;
+    EXPECT_TRUE(run_tasks([&](size_t) -> Status {
+                    ++ran;
+                    return Status::OK();
+                }).ok());
+    EXPECT_EQ(ran, kTaskCount);
+}
+
+TEST_F(SubmitExternalScanTasksTest, ReturnedErrorReachesTheCaller) {
+    const Status result = run_tasks([](size_t idx) -> Status {
+        return idx == kTaskCount - 1 ? Status::InternalError("scanner returned 
an error")
+                                     : Status::OK();
+    });
+    EXPECT_FALSE(result.ok());
+    EXPECT_NE(result.to_string().find("scanner returned an error"), 
std::string::npos);
+}
+
+TEST_F(SubmitExternalScanTasksTest, ThrownExceptionReachesTheCaller) {
+    // The last task is the interesting one: it is the completion that 
releases the
+    // waiter, so a status lost here is a status the caller never sees.
+    const Status result = run_tasks([](size_t idx) -> Status {
+        if (idx == kTaskCount - 1) {
+            throw Exception(ErrorCode::INTERNAL_ERROR, "scanner threw");
+        }
+        return Status::OK();
+    });
+    EXPECT_FALSE(result.ok());
+    EXPECT_NE(result.to_string().find("scanner threw"), std::string::npos);
+}
+
 } // namespace doris
diff --git 
a/regression-test/data/external_table_p0/hive/test_hive_topn_lazy_mat.out 
b/regression-test/data/external_table_p0/hive/test_hive_topn_lazy_mat.out
index 0d4e28fd9f0..c6533162a15 100644
--- a/regression-test/data/external_table_p0/hive/test_hive_topn_lazy_mat.out
+++ b/regression-test/data/external_table_p0/hive/test_hive_topn_lazy_mat.out
@@ -47,6 +47,54 @@
 17.0   user17  17      2
 18.0   user18  18      2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+user4  user4
+user5  user5
+user6  user6
+user7  user7
+user8  user8
+user9  user9
+user10 user10
+
+-- !dup_col_thrice --
+user1  user1   user1
+user2  user2   user2
+user3  user3   user3
+user4  user4   user4
+user5  user5   user5
+user6  user6   user6
+user7  user7   user7
+user8  user8   user8
+user9  user9   user9
+user10 user10  user10
+
+-- !dup_col_mixed --
+1      user1   0.5     user1
+2      user2   1       user2
+3      user3   1.5     user3
+4      user4   2       user4
+5      user5   2.5     user5
+6      user6   3       user6
+7      user7   3.5     user7
+8      user8   4       user8
+9      user9   4.5     user9
+10     user10  5       user10
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+3      3
+3.5    3.5
+4      4
+4.5    4.5
+5      5
+
 -- !test_basic --
 1      user1   1.0     false   0.5     1
 2      user2   2.0     true    1.0     1
@@ -293,6 +341,54 @@
 17.0   user17  17      2
 18.0   user18  18      2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+user4  user4
+user5  user5
+user6  user6
+user7  user7
+user8  user8
+user9  user9
+user10 user10
+
+-- !dup_col_thrice --
+user1  user1   user1
+user2  user2   user2
+user3  user3   user3
+user4  user4   user4
+user5  user5   user5
+user6  user6   user6
+user7  user7   user7
+user8  user8   user8
+user9  user9   user9
+user10 user10  user10
+
+-- !dup_col_mixed --
+1      user1   0.5     user1
+2      user2   1       user2
+3      user3   1.5     user3
+4      user4   2       user4
+5      user5   2.5     user5
+6      user6   3       user6
+7      user7   3.5     user7
+8      user8   4       user8
+9      user9   4.5     user9
+10     user10  5       user10
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+3      3
+3.5    3.5
+4      4
+4.5    4.5
+5      5
+
 -- !test_basic --
 1      user1   1.0     false   0.5     1
 2      user2   2.0     true    1.0     1
@@ -721,6 +817,11 @@
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+
 -- !test_join1 --
 2      user2   2.0     1.0
 4      user4   4.0     2.0
@@ -966,6 +1067,16 @@
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+user4  user4   4       4
+user5  user5   5       5
+user6  user6   6       6
+user7  user7   7       7
+user8  user8   8       8
+
 -- !test_join1 --
 2      user2   2.0     1.0
 4      user4   4.0     2.0
@@ -1240,6 +1351,35 @@
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+user4  user4   4       4
+user5  user5   5       5
+user6  user6   6       6
+user7  user7   7       7
+user8  user8   8       8
+user9  user9   9       9
+user10 user10  10      10
+user11 user11  11      11
+user12 user12  12      12
+user13 user13  13      13
+user14 user14  14      14
+user15 user15  15      15
+user16 user16  16      16
+user17 user17  17      17
+user18 user18  18      18
+user19 user19  19      19
+user20 user20  20      20
+user21 user21  21      21
+user22 user22  22      22
+user23 user23  23      23
+user24 user24  24      24
+user25 user25  25      25
+user26 user26  26      26
+user27 user27  27      27
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1301,6 +1441,31 @@ text_22  {22:["a", null], 23:["b", "c"], 24:["dd", 
null]}        {"a":22, "b":[4, 5]}
 3      text_3  {"a":3, "b":null}       {3:["p", "q"], 4:null, 5:["z1", "z2"]}
 5      text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 6:null, 7:["x1", "x2"]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1407,6 +1572,56 @@ text_17  {17:["t", null], 18:["u1", "u2"], 19:null}      
{"a":null, "b":[17, 18]}
 13     text_13 {"a":13, "b":null}      {13:["d", null], 14:["e1", "e2"], 
15:["f1", null]}
 15     text_15 {"a":null, "b":null}    {15:["z", null], 16:["y1", "y2"], 
17:null}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1597,6 +1812,123 @@ text_1  {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 21     text_21 {"a":21, "b":null}      {21:["z1", "z2"], 22:["m1", null], 
23:["k1", "k2"]}
 23     text_23 {"a":null, "b":[6, 7]}  {23:["k", null], 24:["q1", "q2"], 
25:["tt", null]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+{"a":9, "b":[9, 10]}   {"a":9, "b":[9, 10]}
+{"a":10, "b":null}     {"a":10, "b":null}
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}
+{"a":13, "b":null}     {"a":13, "b":null}
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}
+{"a":null, "b":null}   {"a":null, "b":null}
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}
+{"a":18, "b":null}     {"a":18, "b":null}
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":21, "b":null}     {"a":21, "b":null}
+{"a":22, "b":[4, 5]}   {"a":22, "b":[4, 5]}
+{"a":null, "b":[6, 7]} {"a":null, "b":[6, 7]}
+{"a":24, "b":null}     {"a":24, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+{9:["aa", null], 10:["bb", "cc"], 11:["mm", null]}     {9:["aa", null], 
10:["bb", "cc"], 11:["mm", null]}
+{10:["q1", null], 11:["w1", "w2"], 12:["x", null]}     {10:["q1", null], 
11:["w1", "w2"], 12:["x", null]}
+{11:["x", null], 12:["u1", "u2"], 13:["l", null]}      {11:["x", null], 
12:["u1", "u2"], 13:["l", null]}
+{12:["p", null], 13:["k1", "k2"], 14:null}     {12:["p", null], 13:["k1", 
"k2"], 14:null}
+{13:["d", null], 14:["e1", "e2"], 15:["f1", null]}     {13:["d", null], 
14:["e1", "e2"], 15:["f1", null]}
+{14:["f", null], 15:["g1", "g2"], 16:["h1", null]}     {14:["f", null], 
15:["g1", "g2"], 16:["h1", null]}
+{15:["z", null], 16:["y1", "y2"], 17:null}     {15:["z", null], 16:["y1", 
"y2"], 17:null}
+{16:["i", null], 17:["j1", "j2"], 18:["k", null]}      {16:["i", null], 
17:["j1", "j2"], 18:["k", null]}
+{17:["t", null], 18:["u1", "u2"], 19:null}     {17:["t", null], 18:["u1", 
"u2"], 19:null}
+{18:["v", null], 19:["w1", "w2"], 20:["aa", null]}     {18:["v", null], 
19:["w1", "w2"], 20:["aa", null]}
+{19:["x1", null], 20:["x2", "x3"], 21:["pp", null]}    {19:["x1", null], 
20:["x2", "x3"], 21:["pp", null]}
+{20:["y1", null], 21:null, 22:["qq", "rr"]}    {20:["y1", null], 21:null, 
22:["qq", "rr"]}
+{21:["z1", "z2"], 22:["m1", null], 23:["k1", "k2"]}    {21:["z1", "z2"], 
22:["m1", null], 23:["k1", "k2"]}
+{22:["a", null], 23:["b", "c"], 24:["dd", null]}       {22:["a", null], 
23:["b", "c"], 24:["dd", null]}
+{23:["k", null], 24:["q1", "q2"], 25:["tt", null]}     {23:["k", null], 
24:["q1", "q2"], 25:["tt", null]}
+{24:["u1", null], 25:["u2", "u3"], 26:null}    {24:["u1", null], 25:["u2", 
"u3"], 26:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+text_9 text_9
+text_10        text_10
+text_11        text_11
+text_12        text_12
+text_13        text_13
+text_14        text_14
+text_15        text_15
+text_16        text_16
+text_17        text_17
+text_18        text_18
+text_19        text_19
+text_20        text_20
+text_21        text_21
+text_22        text_22
+text_23        text_23
+text_24        text_24
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+9      {"a":9, "b":[9, 10]}    text_9  {"a":9, "b":[9, 10]}    {9:["aa", 
null], 10:["bb", "cc"], 11:["mm", null]}      text_9
+10     {"a":10, "b":null}      text_10 {"a":10, "b":null}      {10:["q1", 
null], 11:["w1", "w2"], 12:["x", null]}      text_10
+11     {"a":11, "b":[11, 12]}  text_11 {"a":11, "b":[11, 12]}  {11:["x", 
null], 12:["u1", "u2"], 13:["l", null]}       text_11
+12     {"a":null, "b":[12, 13]}        text_12 {"a":null, "b":[12, 13]}        
{12:["p", null], 13:["k1", "k2"], 14:null}      text_12
+13     {"a":13, "b":null}      text_13 {"a":13, "b":null}      {13:["d", 
null], 14:["e1", "e2"], 15:["f1", null]}      text_13
+14     {"a":14, "b":[14, 15]}  text_14 {"a":14, "b":[14, 15]}  {14:["f", 
null], 15:["g1", "g2"], 16:["h1", null]}      text_14
+15     {"a":null, "b":null}    text_15 {"a":null, "b":null}    {15:["z", 
null], 16:["y1", "y2"], 17:null}      text_15
+16     {"a":16, "b":[16, 17]}  text_16 {"a":16, "b":[16, 17]}  {16:["i", 
null], 17:["j1", "j2"], 18:["k", null]}       text_16
+17     {"a":null, "b":[17, 18]}        text_17 {"a":null, "b":[17, 18]}        
{17:["t", null], 18:["u1", "u2"], 19:null}      text_17
+18     {"a":18, "b":null}      text_18 {"a":18, "b":null}      {18:["v", 
null], 19:["w1", "w2"], 20:["aa", null]}      text_18
+19     {"a":19, "b":[1, null]} text_19 {"a":19, "b":[1, null]} {19:["x1", 
null], 20:["x2", "x3"], 21:["pp", null]}     text_19
+20     {"a":null, "b":[2, 3]}  text_20 {"a":null, "b":[2, 3]}  {20:["y1", 
null], 21:null, 22:["qq", "rr"]}     text_20
+21     {"a":21, "b":null}      text_21 {"a":21, "b":null}      {21:["z1", 
"z2"], 22:["m1", null], 23:["k1", "k2"]}     text_21
+22     {"a":22, "b":[4, 5]}    text_22 {"a":22, "b":[4, 5]}    {22:["a", 
null], 23:["b", "c"], 24:["dd", null]}        text_22
+23     {"a":null, "b":[6, 7]}  text_23 {"a":null, "b":[6, 7]}  {23:["k", 
null], 24:["q1", "q2"], 25:["tt", null]}      text_23
+24     {"a":24, "b":null}      text_24 {"a":24, "b":null}      {24:["u1", 
null], 25:["u2", "u3"], 26:null}     text_24
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}        text_17
+{"a":18, "b":null}     {"a":18, "b":null}      text_18
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]} text_19
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1658,6 +1990,31 @@ text_22  {22:["a", null], 23:["b", "c"], 24:["dd", 
null]}        {"a":22, "b":[4, 5]}
 3      text_3  {"a":3, "b":null}       {3:["p", "q"], 4:null, 5:["z1", "z2"]}
 5      text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 6:null, 7:["x1", "x2"]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1764,6 +2121,56 @@ text_17  {17:["t", null], 18:["u1", "u2"], 19:null}      
{"a":null, "b":[17, 18]}
 13     text_13 {"a":13, "b":null}      {13:["d", null], 14:["e1", "e2"], 
15:["f1", null]}
 15     text_15 {"a":null, "b":null}    {15:["z", null], 16:["y1", "y2"], 
17:null}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -1954,6 +2361,123 @@ text_1  {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 21     text_21 {"a":21, "b":null}      {21:["z1", "z2"], 22:["m1", null], 
23:["k1", "k2"]}
 23     text_23 {"a":null, "b":[6, 7]}  {23:["k", null], 24:["q1", "q2"], 
25:["tt", null]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+{"a":9, "b":[9, 10]}   {"a":9, "b":[9, 10]}
+{"a":10, "b":null}     {"a":10, "b":null}
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}
+{"a":13, "b":null}     {"a":13, "b":null}
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}
+{"a":null, "b":null}   {"a":null, "b":null}
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}
+{"a":18, "b":null}     {"a":18, "b":null}
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":21, "b":null}     {"a":21, "b":null}
+{"a":22, "b":[4, 5]}   {"a":22, "b":[4, 5]}
+{"a":null, "b":[6, 7]} {"a":null, "b":[6, 7]}
+{"a":24, "b":null}     {"a":24, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+{9:["aa", null], 10:["bb", "cc"], 11:["mm", null]}     {9:["aa", null], 
10:["bb", "cc"], 11:["mm", null]}
+{10:["q1", null], 11:["w1", "w2"], 12:["x", null]}     {10:["q1", null], 
11:["w1", "w2"], 12:["x", null]}
+{11:["x", null], 12:["u1", "u2"], 13:["l", null]}      {11:["x", null], 
12:["u1", "u2"], 13:["l", null]}
+{12:["p", null], 13:["k1", "k2"], 14:null}     {12:["p", null], 13:["k1", 
"k2"], 14:null}
+{13:["d", null], 14:["e1", "e2"], 15:["f1", null]}     {13:["d", null], 
14:["e1", "e2"], 15:["f1", null]}
+{14:["f", null], 15:["g1", "g2"], 16:["h1", null]}     {14:["f", null], 
15:["g1", "g2"], 16:["h1", null]}
+{15:["z", null], 16:["y1", "y2"], 17:null}     {15:["z", null], 16:["y1", 
"y2"], 17:null}
+{16:["i", null], 17:["j1", "j2"], 18:["k", null]}      {16:["i", null], 
17:["j1", "j2"], 18:["k", null]}
+{17:["t", null], 18:["u1", "u2"], 19:null}     {17:["t", null], 18:["u1", 
"u2"], 19:null}
+{18:["v", null], 19:["w1", "w2"], 20:["aa", null]}     {18:["v", null], 
19:["w1", "w2"], 20:["aa", null]}
+{19:["x1", null], 20:["x2", "x3"], 21:["pp", null]}    {19:["x1", null], 
20:["x2", "x3"], 21:["pp", null]}
+{20:["y1", null], 21:null, 22:["qq", "rr"]}    {20:["y1", null], 21:null, 
22:["qq", "rr"]}
+{21:["z1", "z2"], 22:["m1", null], 23:["k1", "k2"]}    {21:["z1", "z2"], 
22:["m1", null], 23:["k1", "k2"]}
+{22:["a", null], 23:["b", "c"], 24:["dd", null]}       {22:["a", null], 
23:["b", "c"], 24:["dd", null]}
+{23:["k", null], 24:["q1", "q2"], 25:["tt", null]}     {23:["k", null], 
24:["q1", "q2"], 25:["tt", null]}
+{24:["u1", null], 25:["u2", "u3"], 26:null}    {24:["u1", null], 25:["u2", 
"u3"], 26:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+text_9 text_9
+text_10        text_10
+text_11        text_11
+text_12        text_12
+text_13        text_13
+text_14        text_14
+text_15        text_15
+text_16        text_16
+text_17        text_17
+text_18        text_18
+text_19        text_19
+text_20        text_20
+text_21        text_21
+text_22        text_22
+text_23        text_23
+text_24        text_24
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+9      {"a":9, "b":[9, 10]}    text_9  {"a":9, "b":[9, 10]}    {9:["aa", 
null], 10:["bb", "cc"], 11:["mm", null]}      text_9
+10     {"a":10, "b":null}      text_10 {"a":10, "b":null}      {10:["q1", 
null], 11:["w1", "w2"], 12:["x", null]}      text_10
+11     {"a":11, "b":[11, 12]}  text_11 {"a":11, "b":[11, 12]}  {11:["x", 
null], 12:["u1", "u2"], 13:["l", null]}       text_11
+12     {"a":null, "b":[12, 13]}        text_12 {"a":null, "b":[12, 13]}        
{12:["p", null], 13:["k1", "k2"], 14:null}      text_12
+13     {"a":13, "b":null}      text_13 {"a":13, "b":null}      {13:["d", 
null], 14:["e1", "e2"], 15:["f1", null]}      text_13
+14     {"a":14, "b":[14, 15]}  text_14 {"a":14, "b":[14, 15]}  {14:["f", 
null], 15:["g1", "g2"], 16:["h1", null]}      text_14
+15     {"a":null, "b":null}    text_15 {"a":null, "b":null}    {15:["z", 
null], 16:["y1", "y2"], 17:null}      text_15
+16     {"a":16, "b":[16, 17]}  text_16 {"a":16, "b":[16, 17]}  {16:["i", 
null], 17:["j1", "j2"], 18:["k", null]}       text_16
+17     {"a":null, "b":[17, 18]}        text_17 {"a":null, "b":[17, 18]}        
{17:["t", null], 18:["u1", "u2"], 19:null}      text_17
+18     {"a":18, "b":null}      text_18 {"a":18, "b":null}      {18:["v", 
null], 19:["w1", "w2"], 20:["aa", null]}      text_18
+19     {"a":19, "b":[1, null]} text_19 {"a":19, "b":[1, null]} {19:["x1", 
null], 20:["x2", "x3"], 21:["pp", null]}     text_19
+20     {"a":null, "b":[2, 3]}  text_20 {"a":null, "b":[2, 3]}  {20:["y1", 
null], 21:null, 22:["qq", "rr"]}     text_20
+21     {"a":21, "b":null}      text_21 {"a":21, "b":null}      {21:["z1", 
"z2"], 22:["m1", null], 23:["k1", "k2"]}     text_21
+22     {"a":22, "b":[4, 5]}    text_22 {"a":22, "b":[4, 5]}    {22:["a", 
null], 23:["b", "c"], 24:["dd", null]}        text_22
+23     {"a":null, "b":[6, 7]}  text_23 {"a":null, "b":[6, 7]}  {23:["k", 
null], 24:["q1", "q2"], 25:["tt", null]}      text_23
+24     {"a":24, "b":null}      text_24 {"a":24, "b":null}      {24:["u1", 
null], 25:["u2", "u3"], 26:null}     text_24
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}        text_17
+{"a":18, "b":null}     {"a":18, "b":null}      text_18
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]} text_19
+
 -- !1 --
 1      user1   1.0     false   0.5     1
 2      user2   2.0     true    1.0     1
@@ -2002,6 +2526,54 @@ text_1   {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 17.0   user17  17      2
 18.0   user18  18      2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+user4  user4
+user5  user5
+user6  user6
+user7  user7
+user8  user8
+user9  user9
+user10 user10
+
+-- !dup_col_thrice --
+user1  user1   user1
+user2  user2   user2
+user3  user3   user3
+user4  user4   user4
+user5  user5   user5
+user6  user6   user6
+user7  user7   user7
+user8  user8   user8
+user9  user9   user9
+user10 user10  user10
+
+-- !dup_col_mixed --
+1      user1   0.5     user1
+2      user2   1       user2
+3      user3   1.5     user3
+4      user4   2       user4
+5      user5   2.5     user5
+6      user6   3       user6
+7      user7   3.5     user7
+8      user8   4       user8
+9      user9   4.5     user9
+10     user10  5       user10
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+3      3
+3.5    3.5
+4      4
+4.5    4.5
+5      5
+
 -- !test_basic --
 1      user1   1.0     false   0.5     1
 2      user2   2.0     true    1.0     1
@@ -2248,6 +2820,54 @@ text_1   {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 17.0   user17  17      2
 18.0   user18  18      2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+user4  user4
+user5  user5
+user6  user6
+user7  user7
+user8  user8
+user9  user9
+user10 user10
+
+-- !dup_col_thrice --
+user1  user1   user1
+user2  user2   user2
+user3  user3   user3
+user4  user4   user4
+user5  user5   user5
+user6  user6   user6
+user7  user7   user7
+user8  user8   user8
+user9  user9   user9
+user10 user10  user10
+
+-- !dup_col_mixed --
+1      user1   0.5     user1
+2      user2   1       user2
+3      user3   1.5     user3
+4      user4   2       user4
+5      user5   2.5     user5
+6      user6   3       user6
+7      user7   3.5     user7
+8      user8   4       user8
+9      user9   4.5     user9
+10     user10  5       user10
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+3      3
+3.5    3.5
+4      4
+4.5    4.5
+5      5
+
 -- !test_basic --
 1      user1   1.0     false   0.5     1
 2      user2   2.0     true    1.0     1
@@ -2676,6 +3296,11 @@ text_1   {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+
 -- !test_join1 --
 2      user2   2.0     1.0
 4      user4   4.0     2.0
@@ -2921,6 +3546,16 @@ text_1   {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+user4  user4   4       4
+user5  user5   5       5
+user6  user6   6       6
+user7  user7   7       7
+user8  user8   8       8
+
 -- !test_join1 --
 2      user2   2.0     1.0
 4      user4   4.0     2.0
@@ -3195,6 +3830,35 @@ text_1   {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 1      user1   1.0     false   0.5     1       8       user8   8.0     true    
4.0     1
 1      user1   1.0     false   0.5     1       9       user9   9.0     false   
4.5     1
 
+-- !test_join_dup_cols --
+user1  user1   1       1
+user2  user2   2       2
+user3  user3   3       3
+user4  user4   4       4
+user5  user5   5       5
+user6  user6   6       6
+user7  user7   7       7
+user8  user8   8       8
+user9  user9   9       9
+user10 user10  10      10
+user11 user11  11      11
+user12 user12  12      12
+user13 user13  13      13
+user14 user14  14      14
+user15 user15  15      15
+user16 user16  16      16
+user17 user17  17      17
+user18 user18  18      18
+user19 user19  19      19
+user20 user20  20      20
+user21 user21  21      21
+user22 user22  22      22
+user23 user23  23      23
+user24 user24  24      24
+user25 user25  25      25
+user26 user26  26      26
+user27 user27  27      27
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3256,6 +3920,31 @@ text_22  {22:["a", null], 23:["b", "c"], 24:["dd", 
null]}        {"a":22, "b":[4, 5]}
 3      text_3  {"a":3, "b":null}       {3:["p", "q"], 4:null, 5:["z1", "z2"]}
 5      text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 6:null, 7:["x1", "x2"]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3362,6 +4051,56 @@ text_17  {17:["t", null], 18:["u1", "u2"], 19:null}      
{"a":null, "b":[17, 18]}
 13     text_13 {"a":13, "b":null}      {13:["d", null], 14:["e1", "e2"], 
15:["f1", null]}
 15     text_15 {"a":null, "b":null}    {15:["z", null], 16:["y1", "y2"], 
17:null}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3552,6 +4291,123 @@ text_1  {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 21     text_21 {"a":21, "b":null}      {21:["z1", "z2"], 22:["m1", null], 
23:["k1", "k2"]}
 23     text_23 {"a":null, "b":[6, 7]}  {23:["k", null], 24:["q1", "q2"], 
25:["tt", null]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+{"a":9, "b":[9, 10]}   {"a":9, "b":[9, 10]}
+{"a":10, "b":null}     {"a":10, "b":null}
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}
+{"a":13, "b":null}     {"a":13, "b":null}
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}
+{"a":null, "b":null}   {"a":null, "b":null}
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}
+{"a":18, "b":null}     {"a":18, "b":null}
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":21, "b":null}     {"a":21, "b":null}
+{"a":22, "b":[4, 5]}   {"a":22, "b":[4, 5]}
+{"a":null, "b":[6, 7]} {"a":null, "b":[6, 7]}
+{"a":24, "b":null}     {"a":24, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+{9:["aa", null], 10:["bb", "cc"], 11:["mm", null]}     {9:["aa", null], 
10:["bb", "cc"], 11:["mm", null]}
+{10:["q1", null], 11:["w1", "w2"], 12:["x", null]}     {10:["q1", null], 
11:["w1", "w2"], 12:["x", null]}
+{11:["x", null], 12:["u1", "u2"], 13:["l", null]}      {11:["x", null], 
12:["u1", "u2"], 13:["l", null]}
+{12:["p", null], 13:["k1", "k2"], 14:null}     {12:["p", null], 13:["k1", 
"k2"], 14:null}
+{13:["d", null], 14:["e1", "e2"], 15:["f1", null]}     {13:["d", null], 
14:["e1", "e2"], 15:["f1", null]}
+{14:["f", null], 15:["g1", "g2"], 16:["h1", null]}     {14:["f", null], 
15:["g1", "g2"], 16:["h1", null]}
+{15:["z", null], 16:["y1", "y2"], 17:null}     {15:["z", null], 16:["y1", 
"y2"], 17:null}
+{16:["i", null], 17:["j1", "j2"], 18:["k", null]}      {16:["i", null], 
17:["j1", "j2"], 18:["k", null]}
+{17:["t", null], 18:["u1", "u2"], 19:null}     {17:["t", null], 18:["u1", 
"u2"], 19:null}
+{18:["v", null], 19:["w1", "w2"], 20:["aa", null]}     {18:["v", null], 
19:["w1", "w2"], 20:["aa", null]}
+{19:["x1", null], 20:["x2", "x3"], 21:["pp", null]}    {19:["x1", null], 
20:["x2", "x3"], 21:["pp", null]}
+{20:["y1", null], 21:null, 22:["qq", "rr"]}    {20:["y1", null], 21:null, 
22:["qq", "rr"]}
+{21:["z1", "z2"], 22:["m1", null], 23:["k1", "k2"]}    {21:["z1", "z2"], 
22:["m1", null], 23:["k1", "k2"]}
+{22:["a", null], 23:["b", "c"], 24:["dd", null]}       {22:["a", null], 
23:["b", "c"], 24:["dd", null]}
+{23:["k", null], 24:["q1", "q2"], 25:["tt", null]}     {23:["k", null], 
24:["q1", "q2"], 25:["tt", null]}
+{24:["u1", null], 25:["u2", "u3"], 26:null}    {24:["u1", null], 25:["u2", 
"u3"], 26:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+text_9 text_9
+text_10        text_10
+text_11        text_11
+text_12        text_12
+text_13        text_13
+text_14        text_14
+text_15        text_15
+text_16        text_16
+text_17        text_17
+text_18        text_18
+text_19        text_19
+text_20        text_20
+text_21        text_21
+text_22        text_22
+text_23        text_23
+text_24        text_24
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+9      {"a":9, "b":[9, 10]}    text_9  {"a":9, "b":[9, 10]}    {9:["aa", 
null], 10:["bb", "cc"], 11:["mm", null]}      text_9
+10     {"a":10, "b":null}      text_10 {"a":10, "b":null}      {10:["q1", 
null], 11:["w1", "w2"], 12:["x", null]}      text_10
+11     {"a":11, "b":[11, 12]}  text_11 {"a":11, "b":[11, 12]}  {11:["x", 
null], 12:["u1", "u2"], 13:["l", null]}       text_11
+12     {"a":null, "b":[12, 13]}        text_12 {"a":null, "b":[12, 13]}        
{12:["p", null], 13:["k1", "k2"], 14:null}      text_12
+13     {"a":13, "b":null}      text_13 {"a":13, "b":null}      {13:["d", 
null], 14:["e1", "e2"], 15:["f1", null]}      text_13
+14     {"a":14, "b":[14, 15]}  text_14 {"a":14, "b":[14, 15]}  {14:["f", 
null], 15:["g1", "g2"], 16:["h1", null]}      text_14
+15     {"a":null, "b":null}    text_15 {"a":null, "b":null}    {15:["z", 
null], 16:["y1", "y2"], 17:null}      text_15
+16     {"a":16, "b":[16, 17]}  text_16 {"a":16, "b":[16, 17]}  {16:["i", 
null], 17:["j1", "j2"], 18:["k", null]}       text_16
+17     {"a":null, "b":[17, 18]}        text_17 {"a":null, "b":[17, 18]}        
{17:["t", null], 18:["u1", "u2"], 19:null}      text_17
+18     {"a":18, "b":null}      text_18 {"a":18, "b":null}      {18:["v", 
null], 19:["w1", "w2"], 20:["aa", null]}      text_18
+19     {"a":19, "b":[1, null]} text_19 {"a":19, "b":[1, null]} {19:["x1", 
null], 20:["x2", "x3"], 21:["pp", null]}     text_19
+20     {"a":null, "b":[2, 3]}  text_20 {"a":null, "b":[2, 3]}  {20:["y1", 
null], 21:null, 22:["qq", "rr"]}     text_20
+21     {"a":21, "b":null}      text_21 {"a":21, "b":null}      {21:["z1", 
"z2"], 22:["m1", null], 23:["k1", "k2"]}     text_21
+22     {"a":22, "b":[4, 5]}    text_22 {"a":22, "b":[4, 5]}    {22:["a", 
null], 23:["b", "c"], 24:["dd", null]}        text_22
+23     {"a":null, "b":[6, 7]}  text_23 {"a":null, "b":[6, 7]}  {23:["k", 
null], 24:["q1", "q2"], 25:["tt", null]}      text_23
+24     {"a":24, "b":null}      text_24 {"a":24, "b":null}      {24:["u1", 
null], 25:["u2", "u3"], 26:null}     text_24
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}        text_17
+{"a":18, "b":null}     {"a":18, "b":null}      text_18
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]} text_19
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3613,6 +4469,31 @@ text_22  {22:["a", null], 23:["b", "c"], 24:["dd", 
null]}        {"a":22, "b":[4, 5]}
 3      text_3  {"a":3, "b":null}       {3:["p", "q"], 4:null, 5:["z1", "z2"]}
 5      text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 6:null, 7:["x1", "x2"]}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3719,6 +4600,56 @@ text_17  {17:["t", null], 18:["u1", "u2"], 19:null}      
{"a":null, "b":[17, 18]}
 13     text_13 {"a":13, "b":null}      {13:["d", null], 14:["e1", "e2"], 
15:["f1", null]}
 15     text_15 {"a":null, "b":null}    {15:["z", null], 16:["y1", "y2"], 
17:null}
 
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+
 -- !complex_1 --
 1      text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 2:["x", "y"], 3:["k1", 
null]}
 2      text_2  {"a":null, "b":[2, 3]}  {2:["s1", null], 3:["v1", "v2"], 4:null}
@@ -3908,3 +4839,119 @@ text_1  {1:["a", null], 2:["x", "y"], 3:["k1", null]}   
{"a":1, "b":[1, 2]}
 19     text_19 {"a":19, "b":[1, null]} {19:["x1", null], 20:["x2", "x3"], 
21:["pp", null]}
 21     text_21 {"a":21, "b":null}      {21:["z1", "z2"], 22:["m1", null], 
23:["k1", "k2"]}
 23     text_23 {"a":null, "b":[6, 7]}  {23:["k", null], 24:["q1", "q2"], 
25:["tt", null]}
+-- !complex_dup_struct --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":3, "b":null}      {"a":3, "b":null}
+{"a":4, "b":[4, 5]}    {"a":4, "b":[4, 5]}
+{"a":null, "b":[5, 6]} {"a":null, "b":[5, 6]}
+{"a":6, "b":null}      {"a":6, "b":null}
+{"a":7, "b":[7, 8]}    {"a":7, "b":[7, 8]}
+{"a":null, "b":[8, 9]} {"a":null, "b":[8, 9]}
+{"a":9, "b":[9, 10]}   {"a":9, "b":[9, 10]}
+{"a":10, "b":null}     {"a":10, "b":null}
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}
+{"a":13, "b":null}     {"a":13, "b":null}
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}
+{"a":null, "b":null}   {"a":null, "b":null}
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}
+{"a":18, "b":null}     {"a":18, "b":null}
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]}
+{"a":null, "b":[2, 3]} {"a":null, "b":[2, 3]}
+{"a":21, "b":null}     {"a":21, "b":null}
+{"a":22, "b":[4, 5]}   {"a":22, "b":[4, 5]}
+{"a":null, "b":[6, 7]} {"a":null, "b":[6, 7]}
+{"a":24, "b":null}     {"a":24, "b":null}
+
+-- !complex_dup_map --
+{1:["a", null], 2:["x", "y"], 3:["k1", null]}  {1:["a", null], 2:["x", "y"], 
3:["k1", null]}
+{2:["s1", null], 3:["v1", "v2"], 4:null}       {2:["s1", null], 3:["v1", 
"v2"], 4:null}
+{3:["p", "q"], 4:null, 5:["z1", "z2"]} {3:["p", "q"], 4:null, 5:["z1", "z2"]}
+{4:["k1", null], 5:["a1", "a2"], 6:["t1", null]}       {4:["k1", null], 
5:["a1", "a2"], 6:["t1", null]}
+{5:["m", "n"], 6:null, 7:["x1", "x2"]} {5:["m", "n"], 6:null, 7:["x1", "x2"]}
+{6:["z1", null], 7:["t1", "t2"], 8:["c1", null]}       {6:["z1", null], 
7:["t1", "t2"], 8:["c1", null]}
+{7:["h", null], 8:["b1", "b2"], 9:["aa", "bb"]}        {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]}
+{8:["c", null], 9:["d1", "d2"], 10:null}       {8:["c", null], 9:["d1", "d2"], 
10:null}
+{9:["aa", null], 10:["bb", "cc"], 11:["mm", null]}     {9:["aa", null], 
10:["bb", "cc"], 11:["mm", null]}
+{10:["q1", null], 11:["w1", "w2"], 12:["x", null]}     {10:["q1", null], 
11:["w1", "w2"], 12:["x", null]}
+{11:["x", null], 12:["u1", "u2"], 13:["l", null]}      {11:["x", null], 
12:["u1", "u2"], 13:["l", null]}
+{12:["p", null], 13:["k1", "k2"], 14:null}     {12:["p", null], 13:["k1", 
"k2"], 14:null}
+{13:["d", null], 14:["e1", "e2"], 15:["f1", null]}     {13:["d", null], 
14:["e1", "e2"], 15:["f1", null]}
+{14:["f", null], 15:["g1", "g2"], 16:["h1", null]}     {14:["f", null], 
15:["g1", "g2"], 16:["h1", null]}
+{15:["z", null], 16:["y1", "y2"], 17:null}     {15:["z", null], 16:["y1", 
"y2"], 17:null}
+{16:["i", null], 17:["j1", "j2"], 18:["k", null]}      {16:["i", null], 
17:["j1", "j2"], 18:["k", null]}
+{17:["t", null], 18:["u1", "u2"], 19:null}     {17:["t", null], 18:["u1", 
"u2"], 19:null}
+{18:["v", null], 19:["w1", "w2"], 20:["aa", null]}     {18:["v", null], 
19:["w1", "w2"], 20:["aa", null]}
+{19:["x1", null], 20:["x2", "x3"], 21:["pp", null]}    {19:["x1", null], 
20:["x2", "x3"], 21:["pp", null]}
+{20:["y1", null], 21:null, 22:["qq", "rr"]}    {20:["y1", null], 21:null, 
22:["qq", "rr"]}
+{21:["z1", "z2"], 22:["m1", null], 23:["k1", "k2"]}    {21:["z1", "z2"], 
22:["m1", null], 23:["k1", "k2"]}
+{22:["a", null], 23:["b", "c"], 24:["dd", null]}       {22:["a", null], 
23:["b", "c"], 24:["dd", null]}
+{23:["k", null], 24:["q1", "q2"], 25:["tt", null]}     {23:["k", null], 
24:["q1", "q2"], 25:["tt", null]}
+{24:["u1", null], 25:["u2", "u3"], 26:null}    {24:["u1", null], 25:["u2", 
"u3"], 26:null}
+
+-- !complex_dup_string --
+text_1 text_1
+text_2 text_2
+text_3 text_3
+text_4 text_4
+text_5 text_5
+text_6 text_6
+text_7 text_7
+text_8 text_8
+text_9 text_9
+text_10        text_10
+text_11        text_11
+text_12        text_12
+text_13        text_13
+text_14        text_14
+text_15        text_15
+text_16        text_16
+text_17        text_17
+text_18        text_18
+text_19        text_19
+text_20        text_20
+text_21        text_21
+text_22        text_22
+text_23        text_23
+text_24        text_24
+
+-- !complex_dup_mixed --
+1      {"a":1, "b":[1, 2]}     text_1  {"a":1, "b":[1, 2]}     {1:["a", null], 
2:["x", "y"], 3:["k1", null]}   text_1
+2      {"a":null, "b":[2, 3]}  text_2  {"a":null, "b":[2, 3]}  {2:["s1", 
null], 3:["v1", "v2"], 4:null}        text_2
+3      {"a":3, "b":null}       text_3  {"a":3, "b":null}       {3:["p", "q"], 
4:null, 5:["z1", "z2"]}  text_3
+4      {"a":4, "b":[4, 5]}     text_4  {"a":4, "b":[4, 5]}     {4:["k1", 
null], 5:["a1", "a2"], 6:["t1", null]}        text_4
+5      {"a":null, "b":[5, 6]}  text_5  {"a":null, "b":[5, 6]}  {5:["m", "n"], 
6:null, 7:["x1", "x2"]}  text_5
+6      {"a":6, "b":null}       text_6  {"a":6, "b":null}       {6:["z1", 
null], 7:["t1", "t2"], 8:["c1", null]}        text_6
+7      {"a":7, "b":[7, 8]}     text_7  {"a":7, "b":[7, 8]}     {7:["h", null], 
8:["b1", "b2"], 9:["aa", "bb"]} text_7
+8      {"a":null, "b":[8, 9]}  text_8  {"a":null, "b":[8, 9]}  {8:["c", null], 
9:["d1", "d2"], 10:null}        text_8
+9      {"a":9, "b":[9, 10]}    text_9  {"a":9, "b":[9, 10]}    {9:["aa", 
null], 10:["bb", "cc"], 11:["mm", null]}      text_9
+10     {"a":10, "b":null}      text_10 {"a":10, "b":null}      {10:["q1", 
null], 11:["w1", "w2"], 12:["x", null]}      text_10
+11     {"a":11, "b":[11, 12]}  text_11 {"a":11, "b":[11, 12]}  {11:["x", 
null], 12:["u1", "u2"], 13:["l", null]}       text_11
+12     {"a":null, "b":[12, 13]}        text_12 {"a":null, "b":[12, 13]}        
{12:["p", null], 13:["k1", "k2"], 14:null}      text_12
+13     {"a":13, "b":null}      text_13 {"a":13, "b":null}      {13:["d", 
null], 14:["e1", "e2"], 15:["f1", null]}      text_13
+14     {"a":14, "b":[14, 15]}  text_14 {"a":14, "b":[14, 15]}  {14:["f", 
null], 15:["g1", "g2"], 16:["h1", null]}      text_14
+15     {"a":null, "b":null}    text_15 {"a":null, "b":null}    {15:["z", 
null], 16:["y1", "y2"], 17:null}      text_15
+16     {"a":16, "b":[16, 17]}  text_16 {"a":16, "b":[16, 17]}  {16:["i", 
null], 17:["j1", "j2"], 18:["k", null]}       text_16
+17     {"a":null, "b":[17, 18]}        text_17 {"a":null, "b":[17, 18]}        
{17:["t", null], 18:["u1", "u2"], 19:null}      text_17
+18     {"a":18, "b":null}      text_18 {"a":18, "b":null}      {18:["v", 
null], 19:["w1", "w2"], 20:["aa", null]}      text_18
+19     {"a":19, "b":[1, null]} text_19 {"a":19, "b":[1, null]} {19:["x1", 
null], 20:["x2", "x3"], 21:["pp", null]}     text_19
+20     {"a":null, "b":[2, 3]}  text_20 {"a":null, "b":[2, 3]}  {20:["y1", 
null], 21:null, 22:["qq", "rr"]}     text_20
+21     {"a":21, "b":null}      text_21 {"a":21, "b":null}      {21:["z1", 
"z2"], 22:["m1", null], 23:["k1", "k2"]}     text_21
+22     {"a":22, "b":[4, 5]}    text_22 {"a":22, "b":[4, 5]}    {22:["a", 
null], 23:["b", "c"], 24:["dd", null]}        text_22
+23     {"a":null, "b":[6, 7]}  text_23 {"a":null, "b":[6, 7]}  {23:["k", 
null], 24:["q1", "q2"], 25:["tt", null]}      text_23
+24     {"a":24, "b":null}      text_24 {"a":24, "b":null}      {24:["u1", 
null], 25:["u2", "u3"], 26:null}     text_24
+
+-- !complex_dup_pred --
+{"a":1, "b":[1, 2]}    {"a":1, "b":[1, 2]}     text_1
+{"a":10, "b":null}     {"a":10, "b":null}      text_10
+{"a":11, "b":[11, 12]} {"a":11, "b":[11, 12]}  text_11
+{"a":null, "b":[12, 13]}       {"a":null, "b":[12, 13]}        text_12
+{"a":13, "b":null}     {"a":13, "b":null}      text_13
+{"a":14, "b":[14, 15]} {"a":14, "b":[14, 15]}  text_14
+{"a":null, "b":null}   {"a":null, "b":null}    text_15
+{"a":16, "b":[16, 17]} {"a":16, "b":[16, 17]}  text_16
+{"a":null, "b":[17, 18]}       {"a":null, "b":[17, 18]}        text_17
+{"a":18, "b":null}     {"a":18, "b":null}      text_18
+{"a":19, "b":[1, null]}        {"a":19, "b":[1, null]} text_19
diff --git 
a/regression-test/data/external_table_p0/tvf/test_tvf_topn_lazy_mat.out 
b/regression-test/data/external_table_p0/tvf/test_tvf_topn_lazy_mat.out
index a104af3c757..f9647a18b3e 100644
--- a/regression-test/data/external_table_p0/tvf/test_tvf_topn_lazy_mat.out
+++ b/regression-test/data/external_table_p0/tvf/test_tvf_topn_lazy_mat.out
@@ -21,6 +21,24 @@
 1.0    user1   1
 2.0    user2   2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+
+-- !dup_col_mixed --
+1      user1   0.5     user1   0.5
+2      user2   1       user2   1
+3      user3   1.5     user3   1.5
+4      user4   2       user4   2
+
 -- !test_basic --
 1      user1   1.0     false   0.5
 2      user2   2.0     true    1.0
@@ -111,6 +129,24 @@
 1.0    user1   1
 2.0    user2   2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+
+-- !dup_col_mixed --
+1      user1   0.5     user1   0.5
+2      user2   1       user2   1
+3      user3   1.5     user3   1.5
+4      user4   2       user4   2
+
 -- !test_basic --
 1      user1   1.0     false   0.5
 2      user2   2.0     true    1.0
@@ -235,6 +271,24 @@
 1.0    user1   1
 2.0    user2   2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+
+-- !dup_col_mixed --
+1      user1   0.5     user1   0.5
+2      user2   1       user2   1
+3      user3   1.5     user3   1.5
+4      user4   2       user4   2
+
 -- !test_basic --
 1      user1   1.0     false   0.5
 2      user2   2.0     true    1.0
@@ -325,6 +379,24 @@
 1.0    user1   1
 2.0    user2   2
 
+-- !dup_col_twice --
+user1  user1
+user2  user2
+user3  user3
+
+-- !dup_col_nullable --
+0.5    0.5
+1      1
+1.5    1.5
+2      2
+2.5    2.5
+
+-- !dup_col_mixed --
+1      user1   0.5     user1   0.5
+2      user2   1       user2   1
+3      user3   1.5     user3   1.5
+4      user4   2       user4   2
+
 -- !test_basic --
 1      user1   1.0     false   0.5
 2      user2   2.0     true    1.0
diff --git 
a/regression-test/suites/external_table_p0/hive/test_hive_topn_lazy_mat.groovy 
b/regression-test/suites/external_table_p0/hive/test_hive_topn_lazy_mat.groovy
index 222d3a393b2..16e2bd15b55 100644
--- 
a/regression-test/suites/external_table_p0/hive/test_hive_topn_lazy_mat.groovy
+++ 
b/regression-test/suites/external_table_p0/hive/test_hive_topn_lazy_mat.groovy
@@ -34,6 +34,15 @@ suite("test_hive_topn_lazy_mat", 
"p0,external,hive,external_docker,external_dock
             qt_3 """ select score, value, active,name  from ${table} order by 
id,file_id limit 10; """
             qt_4 """ select value,name,id,file_id  from ${table} order by name 
limit 10; """
 
+            // Duplicate projected column - same column twice (the core bug 
from rowid_fetcher fix)
+            qt_dup_col_twice """ select name a, name b from ${table} order by 
id limit 10; """
+            // Duplicate projected column - same column three times
+            qt_dup_col_thrice """ select name a, name b, name c from ${table} 
order by id limit 10; """
+            // Duplicate column mixed with other columns
+            qt_dup_col_mixed """ select id, name a, score, name b from 
${table} order by id limit 10; """
+            // Duplicate nullable column
+            qt_dup_col_nullable """ select score x, score y from ${table} 
order by id limit 10; """
+
 
             for (int limit : limitValues) {
                 // Basic query
@@ -151,10 +160,18 @@ suite("test_hive_topn_lazy_mat", 
"p0,external,hive,external_docker,external_dock
             """
 
             qt_test_join6  """
-                select  * from parquet_topn_lazy_mat_table  as a join 
orc_topn_lazy_mat_table as b 
+                select  * from parquet_topn_lazy_mat_table  as a join 
orc_topn_lazy_mat_table as b
                 where a.file_id = 1  and b.file_id = 1 and a.id = 1
                 order by a.id,b.id  limit 100;
             """
+
+            // Duplicate columns from both sides of join
+            qt_test_join_dup_cols """
+                select a.name, a.name, b.id, b.id
+                from parquet_topn_lazy_mat_table as a
+                join orc_topn_lazy_mat_table as b on a.id = b.id
+                order by a.id limit ${limit};
+            """
         }
 
         for (String table : ["parquet_topn_lazy_complex_table", 
"parquet_topn_lazy_complex_table_multi_pages"]) {
@@ -175,6 +192,15 @@ suite("test_hive_topn_lazy_mat", 
"p0,external,hive,external_docker,external_dock
 
                 qt_complex_12 """ select * from ${table} where id%2 = 0 order 
by id limit ${limit}; """
                 qt_complex_13 """ select * from ${table} where id%2 = 1 order 
by id limit ${limit}; """                
+
+                // Duplicate projected columns over nested types. The scalar 
cases above only
+                // cover text/double/boolean; a STRUCT or MAP column reaches 
the row-id fetch as a
+                // different column shape, so dedup and the result remap need 
coverage here too.
+                qt_complex_dup_struct """ select col2 x, col2 y from ${table} 
order by id limit ${limit}; """
+                qt_complex_dup_map """ select col3 m, col3 n from ${table} 
order by id limit ${limit}; """
+                qt_complex_dup_string """ select col1 a, col1 b from ${table} 
order by id limit ${limit}; """
+                qt_complex_dup_mixed """ select id, col2 a, col1 b, col2 c, 
col3 d, col1 e from ${table} order by id limit ${limit}; """
+                qt_complex_dup_pred """ select col2 x, col2 y, col1 z from 
${table} where col1 like 'text_1%' order by id limit ${limit}; """
             }
         }
     }
@@ -232,6 +258,28 @@ suite("test_hive_topn_lazy_mat", 
"p0,external,hive,external_docker,external_dock
             contains("row_ids: 
[__DORIS_GLOBAL_ROWID_COL__orc_topn_lazy_mat_table]")
         }
 
+        // The duplicate-column regressions below only prove the fix if their 
queries really
+        // enter phase-2 materialization. Pin the plan: two result slots must 
point at one
+        // physical column index, which is the input that used to drop a 
column.
+        explain {
+            sql "select name a, name b from orc_topn_lazy_mat_table order by 
id limit 10;"
+            contains("VMaterializeNode")
+            contains("projectList:[a, b]")
+            contains("column_descs_lists[[`name` text NULL, `name` text 
NULL]]")
+            contains("locations: [[1, 2]]")
+            contains("column_idxs_lists: [[1, 1]]")
+            contains("row_ids: 
[__DORIS_GLOBAL_ROWID_COL__orc_topn_lazy_mat_table]")
+        }
+
+        explain {
+            sql "select col2 x, col2 y from parquet_topn_lazy_complex_table 
order by id limit 10;"
+            contains("VMaterializeNode")
+            contains("column_descs_lists[[`col2` struct<a:int,b:array<int>> 
NULL, `col2` struct<a:int,b:array<int>> NULL]]")
+            contains("locations: [[1, 2]]")
+            contains("column_idxs_lists: [[2, 2]]")
+            contains("row_ids: 
[__DORIS_GLOBAL_ROWID_COL__parquet_topn_lazy_complex_table]")
+        }
+
         explain {
             sql """ select a.name,length(a.name),a.value,b.*,a.* from  
parquet_topn_lazy_mat_table as a    
             join  orc_topn_lazy_mat_table as b on a.id = b.id order by a.name  
  limit 10 """
diff --git 
a/regression-test/suites/external_table_p0/tvf/test_tvf_topn_lazy_mat.groovy 
b/regression-test/suites/external_table_p0/tvf/test_tvf_topn_lazy_mat.groovy
index 6eef30eb939..68c4b952f30 100644
--- a/regression-test/suites/external_table_p0/tvf/test_tvf_topn_lazy_mat.groovy
+++ b/regression-test/suites/external_table_p0/tvf/test_tvf_topn_lazy_mat.groovy
@@ -49,6 +49,13 @@ 
suite("test_tvf_topn_lazy_mat","external,hive,tvf,external_docker") {
             qt_3 """ select score, value, active,name  from ${table} order by 
value limit 4; """
             qt_4 """ select value,name,id from ${table} order by name,score 
limit 2; """
 
+            // Duplicate projected columns: the TVF path shares 
read_batch_external_row() with the
+            // catalog path, so the same physical column projected twice must 
be fetched once and
+            // copied into both result columns.
+            qt_dup_col_twice """ select name a, name b from ${table} order by 
id limit 3; """
+            qt_dup_col_nullable """ select score x, score y from ${table} 
order by id limit 5; """
+            qt_dup_col_mixed """ select id, name a, score, name b, score c 
from ${table} order by id limit 4; """
+
             for (int limit : limitValues) {
                 // Basic query
                 qt_test_basic """ 


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to