dramaticlly commented on code in PR #4717:
URL: https://github.com/apache/iceberg/pull/4717#discussion_r872652195
##########
python/src/iceberg/table/partitioning.py:
##########
@@ -64,3 +67,100 @@ def __str__(self):
def __repr__(self):
return f"PartitionField(field_id={self.field_id}, name={self.name},
transform={repr(self.transform)}, source_id={self.source_id})"
+
+ def __hash__(self):
+ return hash((self.source_id, self.field_id, self.name, self.transform))
+
+
+class PartitionSpec:
+ """
+ PartitionSpec capture the transformation from table data to partition
values
+
+ Attributes:
+ schema(Schema): the schema of data table
+ spec_id(int): any change to PartitionSpec will produce a new specId
+ fields(List[PartitionField): list of partition fields to produce
partition values
+ last_assigned_field_id(int): auto-increment partition field id
starting from PARTITION_DATA_ID_START
+ """
+
+ PARTITION_DATA_ID_START: int = 1000
+
+ def __init__(self, schema: Schema, spec_id: int, fields:
Tuple[PartitionField], last_assigned_field_id: int):
+ self._schema = schema
+ self._spec_id = spec_id
+ self._fields = fields
+ self._last_assigned_field_id = last_assigned_field_id
+ # derived
+ self.fields_by_source_id: Dict[int, List[PartitionField]] = {}
+
+ @property
+ def schema(self) -> Schema:
+ return self._schema
+
+ @property
+ def spec_id(self) -> int:
+ return self._spec_id
+
+ @property
+ def fields(self) -> Tuple[PartitionField]:
+ return self._fields
+
+ @property
+ def last_assigned_field_id(self) -> int:
+ return self._last_assigned_field_id
+
+ def __eq__(self, other):
+ return self.spec_id == other.spec_id and self.fields == other.fields
+
+ def __str__(self):
+ if self.is_unpartitioned():
Review Comment:
>I don't think it is necessary to construct it exactly like Java.
got it, thanks for clarify
> is_unpartitioned can include specs with void transforms
But I am a bit behind on this. From what I can tell, the void transform can
never be unpartitioned per existing java code
https://github.com/apache/iceberg/blob/32841a273b6dda897f6efece4ebb178f8ad54719/api/src/main/java/org/apache/iceberg/PartitionSpec.java#L97-L103.
And if there's one void transform added to such PartitionSpec, then I assume
there's one partition field with voidTransform as its Transform, which is not
`unpartitioned` like below
```
> PartitionSpec ps =
PartitionSpec.builderFor(SCHEMA).alwaysNull("ts").build();
> System.out.println(ps.toString());
[
1000: ts_null: void(5)
]
> System.out.println(ps.isUnpartitioned());
false
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
