This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-doris.git
The following commit(s) were added to refs/heads/master by this push:
new c18717d [fix](vectorized) Fix core dump of mutable block different of
block (#8280)
c18717d is described below
commit c18717df53f25428bcd1bb7b9925b233b520ed64
Author: HappenLee <[email protected]>
AuthorDate: Sat Mar 5 15:27:36 2022 +0800
[fix](vectorized) Fix core dump of mutable block different of block (#8280)
---
be/src/vec/core/block.cpp | 9 ++++++++
be/src/vec/core/block.h | 7 ++----
be/src/vec/exec/join/vhash_join_node.cpp | 39 +++++++++++++++++++-------------
be/src/vec/exec/vset_operation_node.cpp | 6 ++---
4 files changed, 37 insertions(+), 24 deletions(-)
diff --git a/be/src/vec/core/block.cpp b/be/src/vec/core/block.cpp
index 1dda373..23022d3 100644
--- a/be/src/vec/core/block.cpp
+++ b/be/src/vec/core/block.cpp
@@ -835,6 +835,15 @@ doris::Tuple* Block::deep_copy_tuple(const
doris::TupleDescriptor& desc, MemPool
return dst;
}
+MutableBlock::MutableBlock(const std::vector<TupleDescriptor *>& tuple_descs) {
+ for (auto tuple_desc : tuple_descs) {
+ for (auto slot_desc : tuple_desc->slots()) {
+ _data_types.emplace_back(slot_desc->get_data_type_ptr());
+ _columns.emplace_back(_data_types.back()->create_column());
+ }
+ }
+}
+
size_t MutableBlock::rows() const {
for (const auto& column : _columns) {
if (column) {
diff --git a/be/src/vec/core/block.h b/be/src/vec/core/block.h
index 816ed29..c660bdb 100644
--- a/be/src/vec/core/block.h
+++ b/be/src/vec/core/block.h
@@ -306,11 +306,7 @@ public:
MutableBlock() = default;
~MutableBlock() = default;
- MutableBlock(DataTypes data_types) : _columns(data_types.size()),
_data_types(std::move(data_types)) {
- for (int i = 0; i < _data_types.size(); ++i) {
- _columns[i] = _data_types[i]->create_column();
- }
- }
+ MutableBlock(const std::vector<TupleDescriptor*>& tuple_descs);
MutableBlock(Block* block)
: _columns(block->mutate_columns()),
_data_types(block->get_data_types()) {}
@@ -348,6 +344,7 @@ public:
}
}
} else {
+ DCHECK_EQ(_columns.size(), block.columns());
for (int i = 0; i < _columns.size(); ++i) {
if (!_data_types[i]->equals(*block.get_by_position(i).type)) {
DCHECK(_data_types[i]->is_nullable());
diff --git a/be/src/vec/exec/join/vhash_join_node.cpp
b/be/src/vec/exec/join/vhash_join_node.cpp
index f6a5422..9f93f96 100644
--- a/be/src/vec/exec/join/vhash_join_node.cpp
+++ b/be/src/vec/exec/join/vhash_join_node.cpp
@@ -187,11 +187,14 @@ struct ProcessHashTableProbe {
int current_offset = 0;
_items_counts.resize(_probe_rows);
+ _build_block_offsets.resize(_batch_size);
+ _build_block_rows.resize(_batch_size);
memset(_items_counts.data(), 0, sizeof(uint32_t) * _probe_rows);
- constexpr auto is_right_join = JoinOpType::value ==
TJoinOp::RIGHT_ANTI_JOIN ||
+ constexpr auto need_to_set_visited = JoinOpType::value ==
TJoinOp::RIGHT_ANTI_JOIN ||
JoinOpType::value ==
TJoinOp::RIGHT_SEMI_JOIN ||
- JoinOpType::value ==
TJoinOp::RIGHT_OUTER_JOIN;
+ JoinOpType::value ==
TJoinOp::RIGHT_OUTER_JOIN ||
+ JoinOpType::value ==
TJoinOp::FULL_OUTER_JOIN;
constexpr auto is_right_semi_anti_join = JoinOpType::value ==
TJoinOp::RIGHT_ANTI_JOIN ||
JoinOpType::value ==
TJoinOp::RIGHT_SEMI_JOIN;
@@ -230,7 +233,7 @@ struct ProcessHashTableProbe {
// TODO: Iterators are currently considered to be a heavy
operation and have a certain impact on performance.
// We should rethink whether to use this iterator mode in
the future. Now just opt the one row case
if (mapped.get_row_count() == 1) {
- if constexpr (is_right_join)
+ if constexpr (need_to_set_visited)
mapped.visited = true;
if constexpr (!is_right_semi_anti_join) {
@@ -254,7 +257,7 @@ struct ProcessHashTableProbe {
}
++current_offset;
}
- if constexpr (is_right_join)
+ if constexpr (need_to_set_visited)
it->visited = true;
}
}
@@ -294,7 +297,7 @@ struct ProcessHashTableProbe {
mcol[i + right_col_idx]->insert_from(column,
_build_block_rows[j]);
}
} else {
- auto &column =
*_build_blocks[_build_block_offsets[j]].get_by_position(i).column;
+ auto& column =
*_build_blocks[_build_block_offsets[j]].get_by_position(i).column;
mcol[i + right_col_idx]->insert_from(column,
_build_block_rows[j]);
}
}
@@ -369,14 +372,14 @@ struct ProcessHashTableProbe {
for (int i = 0; i < current_offset - origin_offset - 1; ++i) {
same_to_prev.emplace_back(true);
}
- } else if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN
||
+ } else if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN
||
JoinOpType::value == TJoinOp::FULL_OUTER_JOIN
||
JoinOpType::value == TJoinOp::LEFT_ANTI_JOIN)
{
++current_offset;
same_to_prev.emplace_back(false);
visited_map.emplace_back(nullptr);
// only full outer / left outer need insert the data of right
table
- if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
+ if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
JoinOpType::value == TJoinOp::FULL_OUTER_JOIN) {
for (size_t j = 0; j < right_col_len; ++j) {
DCHECK(mcol[j + right_col_idx]->is_nullable());
@@ -414,7 +417,7 @@ struct ProcessHashTableProbe {
(*_join_node->_vother_join_conjunct_ptr)->execute(output_block,
&result_column_id);
auto column =
output_block->get_by_position(result_column_id).column;
- if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
+ if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
JoinOpType::value == TJoinOp::FULL_OUTER_JOIN) {
auto new_filter_column = ColumnVector<UInt8>::create();
auto& filter_map = new_filter_column->get_data();
@@ -492,7 +495,7 @@ struct ProcessHashTableProbe {
output_block->get_by_position(result_column_id).column =
std::move(new_filter_column);
- } else if constexpr (JoinOpType::value == TJoinOp::RIGHT_SEMI_JOIN
||
+ } else if constexpr (JoinOpType::value == TJoinOp::RIGHT_SEMI_JOIN
||
JoinOpType::value ==
TJoinOp::RIGHT_ANTI_JOIN) {
for (int i = 0; i < column->size(); ++i) {
DCHECK(visited_map[i]);
@@ -502,7 +505,7 @@ struct ProcessHashTableProbe {
// inner join do nothing
}
- if constexpr (JoinOpType::value == TJoinOp::RIGHT_SEMI_JOIN ||
+ if constexpr (JoinOpType::value == TJoinOp::RIGHT_SEMI_JOIN ||
JoinOpType::value == TJoinOp::RIGHT_ANTI_JOIN) {
output_block->clear();
} else {
@@ -526,6 +529,7 @@ struct ProcessHashTableProbe {
auto& iter = hash_table_ctx.iter;
auto block_size = 0;
+
auto insert_from_hash_table = [&](uint8_t offset, uint32_t row_num) {
block_size++;
for (size_t j = 0; j < right_col_len; ++j) {
@@ -548,9 +552,9 @@ struct ProcessHashTableProbe {
}
// right outer join / full join need insert data of left table
- if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
- JoinOpType::value == TJoinOp::FULL_OUTER_JOIN ||
- JoinOpType::value == TJoinOp::RIGHT_OUTER_JOIN ||
+ if constexpr (JoinOpType::value == TJoinOp::LEFT_OUTER_JOIN ||
+ JoinOpType::value == TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType::value == TJoinOp::RIGHT_OUTER_JOIN ||
JoinOpType::value == TJoinOp::FULL_OUTER_JOIN) {
for (int i = 0; i < right_col_idx; ++i) {
for (int j = 0; j < block_size; ++j) {
@@ -880,7 +884,7 @@ Status HashJoinNode::open(RuntimeState* state) {
Status HashJoinNode::_hash_table_build(RuntimeState* state) {
RETURN_IF_ERROR(child(1)->open(state));
SCOPED_TIMER(_build_timer);
- MutableBlock mutable_block(_right_table_data_types);
+ MutableBlock mutable_block(child(1)->row_desc().tuple_descriptors());
uint8_t index = 0;
int64_t last_mem_used = 0;
@@ -907,7 +911,7 @@ Status HashJoinNode::_hash_table_build(RuntimeState* state)
{
RETURN_IF_ERROR(_process_build_block(state, _build_blocks[index],
index));
RETURN_IF_LIMIT_EXCEEDED(state, "Hash join, while constructing the
hash table.");
- mutable_block = MutableBlock(_right_table_data_types);
+ mutable_block = MutableBlock();
++index;
last_mem_used = _mem_used;
}
@@ -942,7 +946,10 @@ Status HashJoinNode::extract_build_join_column(Block&
block, NullMap& null_map,
RETURN_IF_ERROR(_build_expr_ctxs[i]->execute(&block,
&result_col_id));
}
- // MutableBlock assume no const column in build block
+ // TODO: opt the column is const
+ block.get_by_position(result_col_id).column =
+
block.get_by_position(result_col_id).column->convert_to_full_column_if_const();
+
if (_is_null_safe_eq_join[i]) {
raw_ptrs[i] = block.get_by_position(result_col_id).column.get();
} else {
diff --git a/be/src/vec/exec/vset_operation_node.cpp
b/be/src/vec/exec/vset_operation_node.cpp
index 9031e0a..63a1ccb 100644
--- a/be/src/vec/exec/vset_operation_node.cpp
+++ b/be/src/vec/exec/vset_operation_node.cpp
@@ -228,13 +228,13 @@ void VSetOperationNode::hash_table_init() {
Status VSetOperationNode::hash_table_build(RuntimeState* state) {
RETURN_IF_ERROR(child(0)->open(state));
Block block;
- MutableBlock mutable_block(_left_table_data_types);
+ MutableBlock mutable_block(child(0)->row_desc().tuple_descriptors());
uint8_t index = 0;
int64_t last_mem_used = 0;
bool eos = false;
while (!eos) {
- block.clear();
+ block.clear_column_data();
SCOPED_TIMER(_build_timer);
RETURN_IF_CANCELLED(state);
RETURN_IF_ERROR(child(0)->get_next(state, &block, &eos));
@@ -254,7 +254,7 @@ Status VSetOperationNode::hash_table_build(RuntimeState*
state) {
// which is better.
RETURN_IF_ERROR(process_build_block(_build_blocks[index], index));
RETURN_IF_LIMIT_EXCEEDED(state, "Set Operation Node, while
constructing the hash table.");
- mutable_block = MutableBlock(_left_table_data_types);
+ mutable_block = MutableBlock();
++index;
last_mem_used = _mem_used;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
