Add ResidualVisitor to compute residuals #1388
Open
+709
−2
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
…ta-only-op add count in data scan and test in catalog sql
kevinjqliu
reviewed
Nov 28, 2024
|
Question: Does it make sense to expose this as the |
…-count Residual Evaluator with test
* added residual evaluator in plan files * tested counts with positional deletes * merged main
* added residual evaluator in plan files * tested counts with positional deletes * merged main * implemented batch reader in count * breaking integration test * fixed integration test * git pull main * revert * revert * revert test_partitioning_key.py * revert test_parser.py * added residual evaluator in visitor * deleted residual_evaluator.py * removed test count from test_sql.py * ignored lint type * fixed lint * working on plan_files * type ignored * make lint
|
Hi @Fokko @kevinjqliu @gli-chris-hao , I have implemented these suggestions with my best understanding.
It would be helpful to get fresh review |
Fokko
changed the title
Fokko
reviewed
Jan 13, 2025
There was a problem hiding this comment.
This is great @tusharchou Thanks for working on this. I left some comments, but this is a great start 🚀
* added residual evaluator in plan files * tested counts with positional deletes * merged main * implemented batch reader in count * breaking integration test * fixed integration test * git pull main * revert * revert * revert test_partitioning_key.py * revert test_parser.py * added residual evaluator in visitor * deleted residual_evaluator.py * removed test count from test_sql.py * ignored lint type * fixed lint * working on plan_files * type ignored * make lint * explicit delete files len is zero * residual eval only if manifest is true * default residual is always true * used projection schema * refactored residual in plan files
* added residual evaluator in plan files * tested counts with positional deletes * merged main * implemented batch reader in count * breaking integration test * fixed integration test * git pull main * revert * revert * revert test_partitioning_key.py * revert test_parser.py * added residual evaluator in visitor * deleted residual_evaluator.py * removed test count from test_sql.py * ignored lint type * fixed lint * working on plan_files * type ignored * make lint * explicit delete files len is zero * residual eval only if manifest is true * default residual is always true * used projection schema * refactored residual in plan files * fixed lint issue with isnan
Fokko
reviewed
Feb 6, 2025
Comment on lines
+1746
to
+1753
| 1. If d > day(a) and d < day(b), the residual is always true | ||
| 2. If d == day(a) and d != day(b), the residual is utc_timestamp >= a | ||
| 3. if d == day(b) and d != day(a), the residual is utc_timestamp <= b | ||
| 4. If d == day(a) == day(b), the residual is utc_timestamp >= a and utc_timestamp <= b | ||
|
|
||
| Partition data is passed using StructLike. Residuals are returned by residualFor(StructLike). | ||
|
|
||
| This class is thread-safe. |
There was a problem hiding this comment.
Suggested change
| 1. If d > day(a) and d < day(b), the residual is always true | |
| 2. If d == day(a) and d != day(b), the residual is utc_timestamp >= a | |
| 3. if d == day(b) and d != day(a), the residual is utc_timestamp <= b | |
| 4. If d == day(a) == day(b), the residual is utc_timestamp >= a and utc_timestamp <= b | |
| Partition data is passed using StructLike. Residuals are returned by residualFor(StructLike). | |
| This class is thread-safe. | |
| 1. If d > day(a) and d < day(b), the residual is always true | |
| 2. If d == day(a) and d != day(b), the residual is utc_timestamp > a | |
| 3. if d == day(b) and d != day(a), the residual is utc_timestamp < b | |
| 4. If d == day(a) == day(b), the residual is utc_timestamp > a and utc_timestamp < b | |
| Partition data is passed using StructLike. Residuals are returned by residualFor(StructLike). |
Fokko
reviewed
Feb 6, 2025
|
|
||
| A residual expression is made by partially evaluating an expression using partition values. | ||
| For example, if a table is partitioned by day(utc_timestamp) and is read with a filter expression | ||
| utc_timestamp >= a and utc_timestamp <= b, then there are 4 possible residuals expressions |
There was a problem hiding this comment.
Suggested change
| utc_timestamp >= a and utc_timestamp <= b, then there are 4 possible residuals expressions | |
| utc_timestamp > a and utc_timestamp < b, then there are 4 possible residuals expressions |
Fokko
reviewed
Feb 6, 2025
| # if the result is not a predicate, then it must be a constant like alwaysTrue or alwaysFalse | ||
| strict_result = bound | ||
|
|
||
| if strict_result is not None and isinstance(strict_result, AlwaysTrue): |
There was a problem hiding this comment.
In Python we can simplify this a bit:
Suggested change
| if strict_result is not None and isinstance(strict_result, AlwaysTrue): | |
| if isinstance(strict_result, AlwaysTrue): |
>>> from pyiceberg.expressions import AlwaysTrue
>>> isinstance(None, AlwaysTrue)
False
Fokko
reviewed
Feb 6, 2025
| strict_result = None | ||
|
|
||
| if strict_projection is not None: | ||
| bound = strict_projection.bind(struct_to_schema(self.spec.partition_type(self.schema))) |
There was a problem hiding this comment.
Suggested change
| bound = strict_projection.bind(struct_to_schema(self.spec.partition_type(self.schema))) | |
| bound = strict_projection.bind(struct_to_schema(self.spec.partition_type(self.schema)), case_sensitive=self.case_sensitive) |
Fokko
reviewed
Feb 6, 2025
| inclusive_projection = part.transform.project(part.name, predicate) | ||
| inclusive_result = None | ||
| if inclusive_projection is not None: | ||
| bound_inclusive = inclusive_projection.bind(struct_to_schema(self.spec.partition_type(self.schema))) |
There was a problem hiding this comment.
Suggested change
| bound_inclusive = inclusive_projection.bind(struct_to_schema(self.spec.partition_type(self.schema))) | |
| bound_inclusive = inclusive_projection.bind(struct_to_schema(self.spec.partition_type(self.schema)), case_sensitive=self.case_sensitive) |
Fokko
reviewed
Feb 6, 2025
| return predicate | ||
|
|
||
| def visit_unbound_predicate(self, predicate: UnboundPredicate[L]) -> BooleanExpression: | ||
| bound = predicate.bind(self.schema, case_sensitive=True) |
There was a problem hiding this comment.
Suggested change
| bound = predicate.bind(self.schema, case_sensitive=True) | |
| bound = predicate.bind(self.schema, case_sensitive=self.case_sensitive) |
Fokko
reviewed
Feb 6, 2025
| # if the result is not a predicate, then it must be a constant like alwaysTrue or | ||
| # alwaysFalse | ||
| inclusive_result = bound_inclusive | ||
| if inclusive_result is not None and isinstance(inclusive_result, AlwaysFalse): |
There was a problem hiding this comment.
Suggested change
| if inclusive_result is not None and isinstance(inclusive_result, AlwaysFalse): | |
| if isinstance(inclusive_result, AlwaysFalse): |
Fokko
reviewed
Feb 6, 2025
| if isinstance(bound, BoundPredicate): | ||
| bound_residual = self.visit_bound_predicate(predicate=bound) | ||
| # if isinstance(bound_residual, BooleanExpression): | ||
| if bound_residual not in (AlwaysFalse(), AlwaysTrue()): |
There was a problem hiding this comment.
Suggested change
| if bound_residual not in (AlwaysFalse(), AlwaysTrue()): | |
| if not isinstance(bound_residual, (AlwaysFalse, AlwaysTrue)): |
Fokko
reviewed
Feb 6, 2025
Comment on lines
+1958
to
+1961
| if len(spec.fields) != 0: | ||
| return ResidualEvaluator(spec=spec, expr=expr, schema=schema, case_sensitive=case_sensitive) | ||
| else: | ||
| return UnpartitionedResidualEvaluator(schema=schema, expr=expr) |
There was a problem hiding this comment.
Just as style thing :)
Suggested change
| if len(spec.fields) != 0: | |
| return ResidualEvaluator(spec=spec, expr=expr, schema=schema, case_sensitive=case_sensitive) | |
| else: | |
| return UnpartitionedResidualEvaluator(schema=schema, expr=expr) | |
| return UnpartitionedResidualEvaluator(schema=schema, expr=expr) if spec.is_unpartitioned() else ResidualEvaluator(spec=spec, expr=expr, schema=schema, case_sensitive=case_sensitive) |
Fokko
reviewed
Feb 6, 2025
Comment on lines
+1683
to
+1693
| if task.file.file_size_in_bytes > 512 * 1024 * 1024: | ||
| target_schema = schema_to_pyarrow(self.projection()) | ||
| batches = arrow_scan.to_record_batches([task]) | ||
| from pyarrow import RecordBatchReader | ||
|
|
||
| reader = RecordBatchReader.from_batches(target_schema, batches) | ||
| for batch in reader: | ||
| res += batch.num_rows | ||
| else: | ||
| tbl = arrow_scan.to_table([task]) | ||
| res += len(tbl) |
There was a problem hiding this comment.
Let's keep it simple for now, I don't think we cover the other case in a test
Suggested change
| if task.file.file_size_in_bytes > 512 * 1024 * 1024: | |
| target_schema = schema_to_pyarrow(self.projection()) | |
| batches = arrow_scan.to_record_batches([task]) | |
| from pyarrow import RecordBatchReader | |
| reader = RecordBatchReader.from_batches(target_schema, batches) | |
| for batch in reader: | |
| res += batch.num_rows | |
| else: | |
| tbl = arrow_scan.to_table([task]) | |
| res += len(tbl) | |
| tbl = arrow_scan.to_table([task]) | |
| res += len(tbl) |
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.
closes issue: Count rows as a metadata-only operation #1223