[fix](inverted index) Fix inverted index for MOR unique table #31051
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
|
Thank you for your contribution to Apache Doris. |
Contributor
Author
|
run buildall |
Contributor
|
clang-tidy review says "All clean, LGTM! 👍" |
|
TeamCity be ut coverage result: |
Contributor
|
PR approved by at least one committer and no changes requested. |
github-actions
bot
added
the
approved
Contributor
|
PR approved by anyone and no changes requested. |
yiguolei
reviewed
Feb 19, 2024
| throw new AnalysisException("index should only be used in columns of DUP_KEYS/UNIQUE_KEYS table" | ||
| + " or key columns of AGG_KEYS table. invalid index: " + indexName); | ||
| } else if (keysType == KeysType.UNIQUE_KEYS && !enableUniqueKeyMergeOnWrite | ||
| && indexType == IndexType.INVERTED && properties != null |
Contributor
There was a problem hiding this comment.
Why check inverted index here? If this is not an inverted index, for example, bloomfilter index (or other index we added in the future) , it is wrong.
Contributor
Author
There was a problem hiding this comment.
I tested and found that index can be used for value columns of mor unique table, since there is a guard for possible wrong result: _should_push_down_value_predicates().
BetaRowsetReader::get_segment_iterators(...) {
// ...
if (_should_push_down_value_predicates()) {
if (_read_context->value_predicates != nullptr) {
_read_options.column_predicates.insert(_read_options.column_predicates.end(),
_read_context->value_predicates->begin(),
_read_context->value_predicates->end());
for (auto pred : *(_read_context->value_predicates)) {
if (_read_options.col_id_to_predicates.count(pred->column_id()) < 1) {
_read_options.col_id_to_predicates.insert(
{pred->column_id(), std::make_shared<AndBlockColumnPredicate>()});
}
auto single_column_block_predicate = new SingleColumnBlockPredicate(pred);
_read_options.col_id_to_predicates[pred->column_id()]->add_column_predicate(
single_column_block_predicate);
}
}
}
// ...
}
bool BetaRowsetReader::_should_push_down_value_predicates() const {
// if unique table with rowset [0-x] or [0-1] [2-y] [...],
// value column predicates can be pushdown on rowset [0-x] or [2-y], [2-y]
// must be compaction, not overlapping and don't have sequence column
return _rowset->keys_type() == UNIQUE_KEYS &&
(((_rowset->start_version() == 0 || _rowset->start_version() == 2) &&
!_rowset->_rowset_meta->is_segments_overlapping() &&
_read_context->sequence_id_idx == -1) ||
_read_context->enable_unique_key_merge_on_write);
}
Contributor
Author
There was a problem hiding this comment.
If _should_push_down_value_predicates() is false, predicate on value column can not be pushed down to storage layer where index is applied. So it's safe to use index on value column. But it's too slow for MATCH query if index is not applied, so do not allow inverted index with parser.
Contributor
Author
|
run buildall |
github-actions
bot
removed
the
approved
Contributor
|
clang-tidy review says "All clean, LGTM! 👍" |
|
TeamCity be ut coverage result: |
TPC-H: Total hot run time: 41313 ms |
TPC-DS: Total hot run time: 177143 ms |
ClickBench: Total hot run time: 30.84 s |
|
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
Contributor
Author
|
run buildall |
Contributor
|
clang-tidy review says "All clean, LGTM! 👍" |
|
TeamCity be ut coverage result: |
TPC-H: Total hot run time: 41298 ms |
TPC-DS: Total hot run time: 178834 ms |
ClickBench: Total hot run time: 31.87 s |
|
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
github-actions
bot
added
the
approved
Contributor
|
PR approved by at least one committer and no changes requested. |
feiniaofeiafei
pushed a commit
to feiniaofeiafei/doris
that referenced
this pull request
Feb 23, 2024
…#31051) * [fix](index) Fix index for none key column of unique mor table (apache#31035) * disable INVERTED index with parser on value columns of MOR unique table * add debug log for test_build_index * add debug log * only do index compaction for dup and mow
xiaokang
added a commit
to xiaokang/doris
that referenced
this pull request
Feb 23, 2024
…#31051) * [fix](index) Fix index for none key column of unique mor table (apache#31035) * disable INVERTED index with parser on value columns of MOR unique table * add debug log for test_build_index * add debug log * only do index compaction for dup and mow
yiguolei
pushed a commit
that referenced
this pull request
Feb 23, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 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.
Proposed changes
Issue Number: close #xxx
Further comments
If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...