[improve](function)Refactor distance function return types to float #55184
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
Contributor
|
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
Contributor
Author
|
run buildall |
uchenily
force-pushed
the
improve-update-type
branch
from
66e5255 to
ba4abda
Compare
Contributor
Author
|
run buildall |
| const auto& offsets2 = *arr2.offsets_ptr; | ||
| const auto& nested_col1 = assert_cast<const ColumnFloat64*>(arr1.nested_col.get()); | ||
| const auto& nested_col2 = assert_cast<const ColumnFloat64*>(arr2.nested_col.get()); | ||
| const auto& nested_col1 = assert_cast<const ColumnType*>(arr1.nested_col.get()); |
Contributor
There was a problem hiding this comment.
what if input column is nullable
Contributor
Author
There was a problem hiding this comment.
Now all distance funcs are changed to PropagateNullable, block data type is FLOAT rather than Nullable(FLOAT), so I removed the null_column.
Contributor
There was a problem hiding this comment.
Now all distance funcs are changed to PropagateNullable, block data type is FLOAT rather than Nullable(FLOAT), so I removed the null_column.
If i am not mistaken, propageteNullable means return null if input is null, so there still could have a nullable input?
Contributor
There was a problem hiding this comment.
I think you can add a regression test.
set debug_skip_fold_constant=true;
select l2_distance([1.0, 2.0, 3.0], [1.0, null, 3.0]);
select l2_distance([1.0, 2.0, 3.0], null);Both of above sql return NULL in current master.
Contributor
There was a problem hiding this comment.
@zhiqiang-hhhh it will be ok. because here we del the use_default_implementation_for_nulls the func in doris will dispose the case
Contributor
There was a problem hiding this comment.
@zhiqiang-hhhh it will be ok. because here we del the
use_default_implementation_for_nullsthe func in doris will dispose the case
ok, i have no problem
TPC-H: Total hot run time: 34012 ms |
TPC-DS: Total hot run time: 185272 ms |
ClickBench: Total hot run time: 32.53 s |
HappenLee
reviewed
Aug 25, 2025
HappenLee
reviewed
Aug 25, 2025
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34416 ms |
TPC-DS: Total hot run time: 185338 ms |
ClickBench: Total hot run time: 32.78 s |
HappenLee
reviewed
Aug 25, 2025
uchenily
force-pushed
the
improve-update-type
branch
from
cb6209c to
2c27b30
Compare
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34094 ms |
TPC-DS: Total hot run time: 184567 ms |
ClickBench: Total hot run time: 32.92 s |
Contributor
|
run buildall |
TPC-H: Total hot run time: 33989 ms |
TPC-DS: Total hot run time: 185107 ms |
ClickBench: Total hot run time: 32.5 s |
Contributor
BE UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34334 ms |
TPC-DS: Total hot run time: 187333 ms |
ClickBench: Total hot run time: 32.67 s |
Contributor
Possible file(s) that should be tracked in LFS detected: 🚨The following file(s) exceeds the file size limit:
Consider using |
Contributor
Author
|
run buildall |
uchenily
force-pushed
the
improve-update-type
branch
from
8600428 to
3c2d11d
Compare
Contributor
Possible file(s) that should be tracked in LFS detected: 🚨The following file(s) exceeds the file size limit:
Consider using |
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34455 ms |
Contributor
BE UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
FE UT Coverage ReportIncrement line coverage |
TPC-DS: Total hot run time: 187671 ms |
ClickBench: Total hot run time: 32.64 s |
Contributor
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
FE Regression && UT Coverage ReportIncrement line coverage |
Contributor
Author
|
run p0 |
Contributor
Author
|
run nonConcurrent |
Contributor
|
PR approved by anyone and no changes requested. |
Contributor
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
Contributor
FE Regression Coverage ReportIncrement line coverage |
1 similar comment
Contributor
FE Regression Coverage ReportIncrement line coverage |
github-actions
bot
added
the
approved
Contributor
|
PR approved by at least one committer and no changes requested. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 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.
What problem does this PR solve?
This pull request standardizes the return type of all vector distance functions to float across the codebase, ensuring consistency and improving performance for vector similarity search operations.
Related PR: #54276
Release note
None
Check List (For Author)
Test
Behavior changed:
The array type in the distance function parameter cannot contain null values, otherwise a runtime error will occur.
If the sum of squares of x or y in cosine_distance is 0, in this case, return distance 2 directly to avoid division by zero.
Check List (For Reviewer who merge this PR)