Skip to content

Commit cbf1afd

Browse files
committed
enhance unit test
1 parent dafb0a4 commit cbf1afd

File tree

2 files changed

+29
-31
lines changed

2 files changed

+29
-31
lines changed

pyiceberg/table/snapshots.py

Lines changed: 1 addition & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -92,24 +92,6 @@ class UpdateMetrics:
9292
added_eq_deletes: int
9393
removed_eq_deletes: int
9494

95-
# def clear() {
96-
# self.added_file_size = 0
97-
# self.removed_file_size = 0
98-
# self.added_data_files = 0
99-
# self.removed_data_files = 0
100-
# self.added_eq_delete_files = 0
101-
# self.removed_eq_delete_files = 0
102-
# self.added_pos_delete_files = 0
103-
# self.removed_pos_delete_files = 0
104-
# self.added_delete_files = 0
105-
# self.removed_delete_files = 0
106-
# self.added_records = 0
107-
# self.deleted_records = 0
108-
# self.added_pos_deletes = 0
109-
# self.removed_pos_deletes = 0
110-
# self.added_eq_deletes = 0
111-
# self.removed_eq_deletes = 0
112-
# }
11395
def __init__(self) -> None:
11496
self.added_file_size = 0
11597
self.removed_file_size = 0
@@ -303,7 +285,7 @@ def remove_file(self, data_file: DataFile, partition_spec: Optional[PartitionSpe
303285
self.metrics.remove_file(data_file)
304286
if getattr(data_file, "partition", None) is not None and len(data_file.partition.record_fields()) != 0:
305287
if partition_spec is None or schema is None:
306-
raise ValueError("add data file with partition but without specifying the partiton_spec and schema")
288+
raise ValueError("remove data file with partition but without specifying the partiton_spec and schema")
307289
self.update_partition_metrics(partition_spec=partition_spec, file=data_file, is_add_file=False, schema=schema)
308290

309291
def update_partition_metrics(self, partition_spec: PartitionSpec, file: DataFile, is_add_file: bool, schema: Schema) -> None:

tests/table/test_snapshots.py

Lines changed: 28 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -156,11 +156,6 @@ def data_file() -> DataFile:
156156
)
157157

158158

159-
@pytest.fixture
160-
def data_file_with_partition() -> DataFile:
161-
return DataFile(content=DataFileContent.DATA, record_count=100, file_size_in_bytes=1234, partition=Record(int_field=1))
162-
163-
164159
def test_snapshot_summary_collector(data_file: DataFile) -> None:
165160
ssc = SnapshotSummaryCollector()
166161

@@ -175,7 +170,9 @@ def test_snapshot_summary_collector(data_file: DataFile) -> None:
175170
}
176171

177172

178-
def test_snapshot_summary_collector_with_partition(data_file_with_partition: DataFile) -> None:
173+
def test_snapshot_summary_collector_with_partition() -> None:
174+
# Given
175+
179176
ssc = SnapshotSummaryCollector()
180177

181178
assert ssc.build() == {}
@@ -185,19 +182,38 @@ def test_snapshot_summary_collector_with_partition(data_file_with_partition: Dat
185182
NestedField(field_id=3, name="int_field", field_type=IntegerType(), required=False),
186183
)
187184
spec = PartitionSpec(PartitionField(source_id=3, field_id=1001, transform=IdentityTransform(), name='int_field'))
185+
data_file_1 = DataFile(content=DataFileContent.DATA, record_count=100, file_size_in_bytes=1234, partition=Record(int_field=1))
186+
data_file_2 = DataFile(content=DataFileContent.DATA, record_count=200, file_size_in_bytes=4321, partition=Record(int_field=2))
187+
# When
188+
ssc.add_file(data_file=data_file_1, schema=schema, partition_spec=spec)
189+
ssc.remove_file(data_file=data_file_1, schema=schema, partition_spec=spec)
190+
ssc.remove_file(data_file=data_file_2, schema=schema, partition_spec=spec)
191+
192+
# Then
193+
assert ssc.build() == {
194+
'added-files-size': '1234',
195+
'removed-files-size': '5555',
196+
'added-data-files': '1',
197+
'deleted-data-files': '2',
198+
'added-records': '100',
199+
'deleted-records': '300',
200+
'changed-partition-count': '2',
201+
}
202+
203+
# When
188204
ssc.set_partition_summary_limit(10)
189-
ssc.add_file(data_file=data_file_with_partition, schema=schema, partition_spec=spec)
190-
ssc.remove_file(data_file=data_file_with_partition, schema=schema, partition_spec=spec)
191205

206+
# Then
192207
assert ssc.build() == {
193208
'added-files-size': '1234',
194-
'removed-files-size': '1234',
209+
'removed-files-size': '5555',
195210
'added-data-files': '1',
196-
'deleted-data-files': '1',
211+
'deleted-data-files': '2',
197212
'added-records': '100',
198-
'deleted-records': '100',
199-
'changed-partition-count': '1',
213+
'deleted-records': '300',
214+
'changed-partition-count': '2',
200215
'partitions.int_field=1': 'added-files-size=1234,removed-files-size=1234,added-data-files=1,deleted-data-files=1,added-records=100,deleted-records=100',
216+
'partitions.int_field=2': 'removed-files-size=4321,deleted-data-files=1,deleted-records=200',
201217
}
202218

203219

0 commit comments

Comments
 (0)