amogh-jahagirdar commented on code in PR #245:
URL: https://github.com/apache/iceberg-python/pull/245#discussion_r1475565567
##########
pyiceberg/table/__init__.py:
##########
@@ -2271,3 +2331,241 @@ def commit(self) -> Snapshot:
)
return snapshot
+
+
+class UpdateSpec:
+ _table: Table
+ _name_to_field: Dict[str, PartitionField] = {}
+ _name_to_added_field: Dict[str, PartitionField] = {}
+ _transform_to_field: Dict[Tuple[int, str], PartitionField] = {}
+ _transform_to_added_field: Dict[Tuple[int, str], PartitionField] = {}
+ _renames: Dict[str, str] = {}
+ _added_time_fields: Dict[int, PartitionField] = {}
+ _case_sensitive: bool
+ _adds: List[PartitionField]
+ _deletes: Set[int]
+ _last_assigned_partition_id: int
+ _transaction: Optional[Transaction]
+
+ def __init__(self, table: Table, transaction: Optional[Transaction] =
None, case_sensitive: bool = True) -> None:
+ self._table = table
+ self._name_to_field = {field.name: field for field in
table.spec().fields}
+ self._name_to_added_field = {}
+ self._transform_to_field = {(field.source_id, repr(field.transform)):
field for field in table.spec().fields}
+ self._transform_to_added_field = {}
+ self._adds = []
+ self._deletes = set()
+ if last_partition_id := table.last_partition_id():
+ self._last_assigned_partition_id = last_partition_id
+ else:
+ self._last_assigned_partition_id = PARTITION_FIELD_ID_START - 1
+ self._renames = {}
+ self._transaction = transaction
+ self._case_sensitive = case_sensitive
+ self._added_time_fields = {}
+
+ def add_field(
+ self,
+ source_column_name: str,
+ transform: Transform[Any, Any],
+ partition_field_name: Optional[str] = None,
+ ) -> UpdateSpec:
+ ref = Reference(source_column_name)
+ bound_ref = ref.bind(self._table.schema(), self._case_sensitive)
+ # verify transform can actually bind it
+ output_type = bound_ref.field.field_type
+ if not transform.can_transform(output_type):
+ raise ValueError(f"{transform} cannot transform {output_type}
values from {bound_ref.field.name}")
+
+ transform_key = (bound_ref.field.field_id, repr(transform))
+ existing_partition_field = self._transform_to_field.get(transform_key)
+ if existing_partition_field and
self._is_duplicate_partition(transform, existing_partition_field):
+ raise ValueError(f"Duplicate partition field for
${ref.name}=${ref}, ${existing_partition_field} already exists")
+
+ added = self._transform_to_added_field.get(transform_key)
+ if added:
+ raise ValueError(f"Already added partition: {added.name}")
+
+ new_field = self._partition_field((bound_ref.field.field_id,
transform), partition_field_name)
+ if new_field.name in self._name_to_added_field:
+ raise ValueError(f"Already added partition field with name:
{new_field.name}")
+
+ self._redundant_time_partition(new_field)
+ self._transform_to_added_field[transform_key] = new_field
+
+ existing_partition_field = self._name_to_field.get(new_field.name)
+ if existing_partition_field and new_field.field_id not in
self._deletes:
+ if isinstance(existing_partition_field.transform, VoidTransform):
+ self.rename_field(
+ existing_partition_field.name,
existing_partition_field.name + "_" + str(existing_partition_field.field_id)
+ )
+ else:
+ raise ValueError(f"Cannot add duplicate partition field name:
{existing_partition_field.name}")
+
+ self._name_to_added_field[new_field.name] = new_field
+ self._adds.append(new_field)
+ return self
+
+ def add_identity(self, source_column_name: str) -> UpdateSpec:
+ return self.add_field(source_column_name, IdentityTransform(), None)
+
+ def remove_field(self, name: str) -> UpdateSpec:
+ added = self._name_to_added_field.get(name)
+ if added:
+ raise ValueError(f"Cannot delete newly added field {name}")
+ renamed = self._renames.get(name)
+ if renamed:
+ raise ValueError(f"Cannot rename and delete field {name}")
+ field = self._name_to_field.get(name)
+ if not field:
+ raise ValueError(f"No such partition field: {name}")
+
+ self._deletes.add(field.field_id)
+ return self
+
+ def rename_field(self, name: str, new_name: str) -> UpdateSpec:
+ existing_field = self._name_to_field.get(new_name)
+ if existing_field and isinstance(existing_field.transform,
VoidTransform):
+ return self.rename_field(name, name + "_" +
str(existing_field.field_id))
+ added = self._name_to_added_field.get(name)
+ if added:
+ raise ValueError("Cannot rename recently added partitions")
+ field = self._name_to_field.get(name)
+ if not field:
+ raise ValueError(f"Cannot find partition field {name}")
+ if field.field_id in self._deletes:
+ raise ValueError(f"Cannot delete and rename partition field
{name}")
+ self._renames[name] = new_name
+ return self
+
+ def commit(self) -> None:
+ new_spec = self._apply()
+ if self._table.metadata.default_spec_id != new_spec.spec_id:
+ if new_spec.spec_id not in self._table.specs():
+ updates = [AddPartitionSpecUpdate(spec=new_spec),
SetDefaultSpecUpdate(spec_id=-1)]
Review Comment:
@HonahX setting the -1 spec id for the spec update down here. Let me know
what you think
--
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: issues-unsubscr...@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For additional commands, e-mail: issues-h...@iceberg.apache.org
