[Data] Use sampled fragments to estimate Parquet reader batch size #45749
Conversation
This file contains 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
Signed-off-by: Balaji Veeramani <[email protected]>
Signed-off-by: Balaji Veeramani <[email protected]>
raulchen
approved these changes
Jun 5, 2024
| sample_info.actual_bytes_per_row is None | ||
| or sample_info.estimated_bytes_per_row is None | ||
| ): | ||
| return PARQUET_ENCODING_RATIO_ESTIMATE_LOWER_BOUND |
There was a problem hiding this comment.
if some samples are None while some are not, should we drop the None values? I.E., only fallback to this lower bound if all samples are None.
There was a problem hiding this comment.
I matched the behavior of the existing implementation, but that makes sense to me.
@c21 wdyt (I think you wrote the original implementation)? I can address this in a follow-up PR
ryanaoleary
pushed a commit
to ryanaoleary/ray
that referenced
this pull request
Jun 6, 2024
…ay-project#45749) ParquetDatasource uses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata, ParquetDatasource uses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size. Signed-off-by: Balaji Veeramani <[email protected]> Signed-off-by: Ryan O'Leary <[email protected]>
yucai
pushed a commit
to yucai/ray
that referenced
this pull request
Jun 6, 2024
…ay-project#45749) ParquetDatasource uses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata, ParquetDatasource uses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size. Signed-off-by: Balaji Veeramani <[email protected]> Signed-off-by: yucai <[email protected]>
ryanaoleary
pushed a commit
to ryanaoleary/ray
that referenced
this pull request
Jun 6, 2024
…ay-project#45749) ParquetDatasource uses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata, ParquetDatasource uses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size. Signed-off-by: Balaji Veeramani <[email protected]> Signed-off-by: Ryan O'Leary <[email protected]>
yucai
pushed a commit
to yucai/ray
that referenced
this pull request
Jun 7, 2024
…ay-project#45749) ParquetDatasource uses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata, ParquetDatasource uses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size. Signed-off-by: Balaji Veeramani <[email protected]> Signed-off-by: yucai <[email protected]>
richardsliu
pushed a commit
to richardsliu/ray
that referenced
this pull request
Jun 12, 2024
…ay-project#45749) ParquetDatasource uses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata, ParquetDatasource uses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size. Signed-off-by: Balaji Veeramani <[email protected]> Signed-off-by: Richard Liu <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Why are these changes needed?
ParquetDatasourceuses fetched metadata to determine the Parquet reader batch size. If the metadata provider doesn't provide metadata,ParquetDatasourceuses a default value of 10,000 rows. While this value might be reasonable for tabular data, the default value can lead to poor performance or even errors if each row is large. To avoid this issue, this PR updates the implementation to use the sampled fragments to determine the batch size.Related issue number
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/under thecorresponding
.rstfile.