GH-49069: [C++] Share Trie instances across CSV value decoders #49070
+66
−44
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
|
kou
approved these changes
Jan 30, 2026
cpp/src/arrow/csv/converter.cc
Outdated
| const ConvertOptions& options) | ||
| : type_(type), options_(options) {} | ||
| const ConvertOptions& options, const TrieCache* trie_cache) | ||
| : type_(type), options_(options), trie_cache_(*trie_cache) {} |
Member
There was a problem hiding this comment.
How about using trie_cache_ as a pointer not a reference?
cpp/src/arrow/csv/converter.h
Outdated
| const ConvertOptions& options_; | ||
| MemoryPool* pool_; | ||
| std::shared_ptr<DataType> type_; | ||
| std::shared_ptr<void> trie_cache_; // Opaque pointer to TrieCache |
Member
There was a problem hiding this comment.
TrieCache's destructor is still called with this, right?
Member
Author
There was a problem hiding this comment.
Yes. Let me add a comment.
Member
Author
There was a problem hiding this comment.
Just to confirm, I checked:
The destructor of shared_ptr decrements the number of shared owners of the control block. If that counter reaches zero, the control block calls the destructor of the managed object. The control block does not deallocate itself until the std::weak_ptr counter reaches zero as well.
github-actions
bot
added
awaiting merge
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Rationale for this change
The CSV converter was building identical Trie data structures (for null/true/false values) in every decoder instance, causing duplicate memory allocation and initialization overhead.
What changes are included in this PR?
TrieCachestruct to hold shared Trie instances (null_trie, true_trie, false_trie)ValueDecoderand all decoder subclasses to accept and reference a sharedTrieCacheinstead of building their own TriesConverterbase class to create oneTrieCacheper converter and pass it to all decodersAre these changes tested?
Yes, all existing tests. I ran a simple benchmark showing roughly 2-4% faster converter creation, and obviously less memory usage.
Are there any user-facing changes?
No.