This is an automated email from the ASF dual-hosted git repository.
sungwy pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/iceberg-python.git
The following commit(s) were added to refs/heads/main by this push:
new 40357476 Bug Fix: Use historical partition field name (#1161)
40357476 is described below
commit 40357476ad0a79f7486b96a6e29b404bc699b70d
Author: Sung Yun <[email protected]>
AuthorDate: Thu Sep 12 08:16:33 2024 -0400
Bug Fix: Use historical partition field name (#1161)
* use historical partition field name
* thanks ndrluis!
---
pyiceberg/table/update/spec.py | 10 ++++-----
tests/integration/test_partition_evolution.py | 32 +++++++++++++++++++++++++++
2 files changed, 37 insertions(+), 5 deletions(-)
diff --git a/pyiceberg/table/update/spec.py b/pyiceberg/table/update/spec.py
index aaa0e138..b732b211 100644
--- a/pyiceberg/table/update/spec.py
+++ b/pyiceberg/table/update/spec.py
@@ -275,12 +275,12 @@ class UpdateSpec(UpdateTableMetadata["UpdateSpec"]):
historical_fields = []
for spec in self._transaction.table_metadata.specs().values():
for field in spec.fields:
- historical_fields.append((field.source_id, field.field_id,
repr(field.transform), field.name))
+ historical_fields.append(field)
- for field_key in historical_fields:
- if field_key[0] == source_id and field_key[2] ==
repr(transform):
- if name is None or field_key[3] == name:
- return PartitionField(source_id, field_key[1],
transform, name)
+ for field in historical_fields:
+ if field.source_id == source_id and repr(field.transform) ==
repr(transform):
+ if name is None or field.name == name:
+ return PartitionField(source_id, field.field_id,
transform, field.name)
new_field_id = self._new_field_id()
if name is None:
diff --git a/tests/integration/test_partition_evolution.py
b/tests/integration/test_partition_evolution.py
index 5cc7512f..805c0c1f 100644
--- a/tests/integration/test_partition_evolution.py
+++ b/tests/integration/test_partition_evolution.py
@@ -207,6 +207,38 @@ def test_remove_identity_v2(catalog: Catalog) -> None:
assert table_v2.spec() == PartitionSpec(spec_id=0)
[email protected]
[email protected]("catalog",
[pytest.lazy_fixture("session_catalog_hive"),
pytest.lazy_fixture("session_catalog")])
+def test_remove_and_add_identity(catalog: Catalog) -> None:
+ table = _table(catalog)
+ table.update_spec().add_identity("id").commit()
+ table.update_spec().remove_field("id").commit()
+ table.update_spec().add_identity("id").commit()
+
+ assert len(table.specs()) == 4
+ assert table.spec().spec_id == 3
+ assert table.spec() == PartitionSpec(
+ PartitionField(source_id=1, field_id=1000, transform=VoidTransform(),
name="id_1000"),
+ PartitionField(source_id=1, field_id=1001,
transform=IdentityTransform(), name="id"),
+ spec_id=3,
+ )
+
+
[email protected]
[email protected]("catalog",
[pytest.lazy_fixture("session_catalog_hive"),
pytest.lazy_fixture("session_catalog")])
+def test_remove_and_add_identity_v2(catalog: Catalog) -> None:
+ table_v2 = _table_v2(catalog)
+ table_v2.update_spec().add_identity("id").commit()
+ table_v2.update_spec().remove_field("id").commit()
+ table_v2.update_spec().add_identity("id").commit()
+
+ assert len(table_v2.specs()) == 2
+ assert table_v2.spec().spec_id == 1
+ assert table_v2.spec() == PartitionSpec(
+ PartitionField(source_id=1, field_id=1000,
transform=IdentityTransform(), name="id"), spec_id=1
+ )
+
+
@pytest.mark.integration
@pytest.mark.parametrize("catalog",
[pytest.lazy_fixture("session_catalog_hive"),
pytest.lazy_fixture("session_catalog")])
def test_remove_bucket(catalog: Catalog) -> None:
