rok commented on code in PR #50945:
URL: https://github.com/apache/arrow/pull/50945#discussion_r3914332413


##########
cpp/src/arrow/json/chunker.cc:
##########
@@ -124,49 +56,117 @@ namespace {
 // and uses actual JSON parsing to delimit them.
 class ParsingBoundaryFinder : public BoundaryFinder {
  public:
-  Status FindFirst(string_view partial, string_view block, int64_t* out_pos) 
override {
-    auto length = ConsumeWholeObject(MultiStringStream({partial, block}));
-    if (length == string_view::npos) {
+  Status FindFirst(std::string_view partial, std::string_view block,
+                   int64_t* out_pos) override {
+    simdjson::padded_string input;
+
+    if (partial.empty()) {
+      input = simdjson::padded_string(block);
+    } else if (block.empty()) {
+      input = simdjson::padded_string(partial);
+    } else {
+      simdjson::padded_string_builder builder(partial.size() + block.size());
+      builder.append(partial);
+      builder.append(block);
+      input = builder.convert();
+    }
+
+    const auto length = ConsumeWholeObject(input);
+
+    if (length == std::string_view::npos) {
       *out_pos = -1;
     } else if (ARROW_PREDICT_FALSE(length < partial.size())) {
       return Status::Invalid("JSON chunk error: invalid data at end of 
document");
     } else {
       DCHECK_LE(length, partial.size() + block.size());
       *out_pos = static_cast<int64_t>(length - partial.size());
     }
+
     return Status::OK();
   }
 
   Status FindLast(std::string_view block, int64_t* out_pos) override {
-    const size_t block_length = block.size();
     size_t consumed_length = 0;
-    while (consumed_length < block_length) {
-      rj::MemoryStream ms(reinterpret_cast<const char*>(block.data()), 
block.size());
-      using InputStream = rj::EncodedInputStream<rj::UTF8<>, rj::MemoryStream>;
-      auto length = ConsumeWholeObject(InputStream(ms));
-      if (length == string_view::npos || length == 0) {
-        // found incomplete object or block is empty
+
+    simdjson::ondemand::document_stream stream;
+
+    if (parser_.iterate_many(GetPaddedStringView(block)).get(stream) !=
+        simdjson::SUCCESS) {
+      *out_pos = -1;
+      return Status::OK();
+    }
+
+    auto it = stream.begin();
+    if (it == stream.end()) {
+      *out_pos = -1;
+      return Status::OK();
+    }
+
+    while (it != stream.end()) {
+      if (!ConsumeDocument(it).ok()) {
         break;
       }
-      consumed_length += length;
-      block = block.substr(length);
+      consumed_length = it.current_index() + it.source().size();
+      ++it;
     }
+
     if (consumed_length == 0) {
       *out_pos = -1;
     } else {
-      consumed_length += ConsumeWhitespace(block);
-      DCHECK_LE(consumed_length, block_length);
+      consumed_length += ConsumeWhitespace(block.substr(consumed_length));
       *out_pos = static_cast<int64_t>(consumed_length);
     }
+
     return Status::OK();
   }
-
   Status FindNth(std::string_view partial, std::string_view block, int64_t 
count,
                  int64_t* out_pos, int64_t* num_found) override {
     return Status::NotImplemented("ParsingBoundaryFinder::FindNth");
   }
-};
 
+ private:
+  simdjson::ondemand::parser parser_;
+  std::string buffer_;
+
+  simdjson::padded_string_view GetPaddedStringView(std::string_view input) {
+    buffer_.assign(input);
+    if (buffer_.capacity() - buffer_.size() < simdjson::SIMDJSON_PADDING) {
+      buffer_.reserve(buffer_.size() + simdjson::SIMDJSON_PADDING);
+    }
+    return simdjson::padded_string_view(buffer_.data(), buffer_.size(),
+                                        buffer_.capacity());
+  }
+
+  size_t ConsumeWholeObject(std::string_view input) {
+    if (input.size() == 0) {
+      return 0;
+    }
+
+    simdjson::ondemand::document_stream stream;
+
+    if (parser_.iterate_many(GetPaddedStringView(input)).get(stream) !=

Review Comment:
   See comment above
   ```suggestion
       if (parser_.iterate_many(GetPaddedStringView(input), 
input.size()).get(stream) !=
   ```



##########
cpp/src/arrow/json/chunker.cc:
##########
@@ -124,49 +56,117 @@ namespace {
 // and uses actual JSON parsing to delimit them.
 class ParsingBoundaryFinder : public BoundaryFinder {
  public:
-  Status FindFirst(string_view partial, string_view block, int64_t* out_pos) 
override {
-    auto length = ConsumeWholeObject(MultiStringStream({partial, block}));
-    if (length == string_view::npos) {
+  Status FindFirst(std::string_view partial, std::string_view block,
+                   int64_t* out_pos) override {
+    simdjson::padded_string input;
+
+    if (partial.empty()) {
+      input = simdjson::padded_string(block);
+    } else if (block.empty()) {
+      input = simdjson::padded_string(partial);
+    } else {
+      simdjson::padded_string_builder builder(partial.size() + block.size());
+      builder.append(partial);
+      builder.append(block);
+      input = builder.convert();
+    }
+
+    const auto length = ConsumeWholeObject(input);
+
+    if (length == std::string_view::npos) {
       *out_pos = -1;
     } else if (ARROW_PREDICT_FALSE(length < partial.size())) {
       return Status::Invalid("JSON chunk error: invalid data at end of 
document");
     } else {
       DCHECK_LE(length, partial.size() + block.size());
       *out_pos = static_cast<int64_t>(length - partial.size());
     }
+
     return Status::OK();
   }
 
   Status FindLast(std::string_view block, int64_t* out_pos) override {
-    const size_t block_length = block.size();
     size_t consumed_length = 0;
-    while (consumed_length < block_length) {
-      rj::MemoryStream ms(reinterpret_cast<const char*>(block.data()), 
block.size());
-      using InputStream = rj::EncodedInputStream<rj::UTF8<>, rj::MemoryStream>;
-      auto length = ConsumeWholeObject(InputStream(ms));
-      if (length == string_view::npos || length == 0) {
-        // found incomplete object or block is empty
+
+    simdjson::ondemand::document_stream stream;
+
+    if (parser_.iterate_many(GetPaddedStringView(block)).get(stream) !=

Review Comment:
   `iterate_many` [rejects strings larger than 
1mb](https://github.com/simdjson/simdjson/blob/v4.6.4/include/simdjson/generic/ondemand/parser.h#L215-L216).
 We could:
   
   ```suggestion
       if (parser_.iterate_many(GetPaddedStringView(block), 
block.size()).get(stream) !=
   ```



##########
cpp/src/arrow/json/chunker.cc:
##########
@@ -124,49 +56,117 @@ namespace {
 // and uses actual JSON parsing to delimit them.
 class ParsingBoundaryFinder : public BoundaryFinder {
  public:
-  Status FindFirst(string_view partial, string_view block, int64_t* out_pos) 
override {
-    auto length = ConsumeWholeObject(MultiStringStream({partial, block}));
-    if (length == string_view::npos) {
+  Status FindFirst(std::string_view partial, std::string_view block,
+                   int64_t* out_pos) override {
+    simdjson::padded_string input;
+
+    if (partial.empty()) {
+      input = simdjson::padded_string(block);
+    } else if (block.empty()) {
+      input = simdjson::padded_string(partial);
+    } else {
+      simdjson::padded_string_builder builder(partial.size() + block.size());

Review Comment:
   `padded_string_builder` was introduced in 
[v4.3.0](https://github.com/simdjson/simdjson/releases/tag/v4.3.0) so builds 
against 4.0-4.2 will likely fail.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to