no need to copy strings to a heap, we have already copied the data to…

… the segment, the map is only ever referenced while the segment is alive, so we can safely reference data from the segment
Tishj · Jan 7, 2025 · d5aea48 · d5aea48
1 parent 12d1051
commit d5aea48
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/src/include/duckdb/storage/compression/dictionary/compression.hpp b/src/include/duckdb/storage/compression/dictionary/compression.hpp
@@ -47,7 +47,6 @@ struct DictionaryCompressionCompressState : public DictionaryCompressionState {
 	data_ptr_t current_end_ptr;
 
 	// Buffers and map for current segment
-	StringHeap heap;
 	string_map_t<uint32_t> current_string_map;
 	vector<uint32_t> index_buffer;
 	vector<uint32_t> selection_buffer;

diff --git a/src/storage/compression/dictionary/compression.cpp b/src/storage/compression/dictionary/compression.cpp
@@ -6,8 +6,7 @@ namespace duckdb {
 DictionaryCompressionCompressState::DictionaryCompressionCompressState(ColumnDataCheckpointData &checkpoint_data_p,
                                                                        const CompressionInfo &info)
     : DictionaryCompressionState(info), checkpoint_data(checkpoint_data_p),
-      function(checkpoint_data.GetCompressionFunction(CompressionType::COMPRESSION_DICTIONARY)),
-      heap(BufferAllocator::Get(checkpoint_data.GetDatabase())) {
+      function(checkpoint_data.GetCompressionFunction(CompressionType::COMPRESSION_DICTIONARY)) {
 	CreateEmptySegment(checkpoint_data.GetRowGroup().start);
 }
 
@@ -72,7 +71,9 @@ void DictionaryCompressionCompressState::AddNewString(string_t str) {
 	if (str.IsInlined()) {
 		current_string_map.insert({str, index_buffer.size() - 1});
 	} else {
-		current_string_map.insert({heap.AddBlob(str), index_buffer.size() - 1});
+		string_t dictionary_string((const char *)dict_pos, UnsafeNumericCast<uint32_t>(str.GetSize())); // NOLINT
+		D_ASSERT(!dictionary_string.IsInlined());
+		current_string_map.insert({dictionary_string, index_buffer.size() - 1});
 	}
 	DictionaryCompression::SetDictionary(*current_segment, current_handle, current_dictionary);