fixup! Add table statistics update

apache · Nov 12, 2024 · 11120bf · 11120bf
1 parent d3aaab3
commit 11120bf
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 6 deletions.
diff --git a/mkdocs/docs/api.md b/mkdocs/docs/api.md
@@ -1140,15 +1140,16 @@ table.update_statistics().set_statistics(snapshot_id, statistics_file).commit()
 table.update_statistics()
   .set_statistics(snapshot_id1, statistics_file1)
   .remove_statistics(snapshot_id2)
+  .commit()
 # Operations are applied on commit.
 ```
 
 You can also use context managers to make more changes:
 
 ```python
 with table.update_statistics() as update:
-  update.set_statistics(1, statistics_file)
-  update.remove_statistics(2)
+  update.set_statistics(snaphsot_id1, statistics_file)
+  update.remove_statistics(snapshot_id2)
 ```
 
 ## Query the data

diff --git a/pyiceberg/table/statistics.py b/pyiceberg/table/statistics.py
@@ -42,7 +42,7 @@ class StatisticsFile(IcebergBaseModel):
     blob_metadata: List[BlobMetadata] = Field(alias="blob-metadata")
 
 
-def reject_statistics(
+def filter_statistics_by_snapshot_id(
     statistics: List[StatisticsFile],
     reject_snapshot_id: int,
 ) -> List[StatisticsFile]:

diff --git a/pyiceberg/table/update/__init__.py b/pyiceberg/table/update/__init__.py
@@ -37,7 +37,7 @@
     SnapshotLogEntry,
 )
 from pyiceberg.table.sorting import SortOrder
-from pyiceberg.table.statistics import StatisticsFile, reject_statistics
+from pyiceberg.table.statistics import StatisticsFile, filter_statistics_by_snapshot_id
 from pyiceberg.typedef import (
     IcebergBaseModel,
     Properties,
@@ -496,7 +496,7 @@ def _(update: SetStatisticsUpdate, base_metadata: TableMetadata, context: _Table
     if update.snapshot_id != update.statistics.snapshot_id:
         raise ValueError("Snapshot id in statistics does not match the snapshot id in the update")
 
-    statistics = reject_statistics(base_metadata.statistics, update.snapshot_id)
+    statistics = filter_statistics_by_snapshot_id(base_metadata.statistics, update.snapshot_id)
     context.add_update(update)
 
     return base_metadata.model_copy(update={"statistics": statistics + [update.statistics]})
@@ -507,7 +507,7 @@ def _(update: RemoveStatisticsUpdate, base_metadata: TableMetadata, context: _Ta
     if not any(stat.snapshot_id == update.snapshot_id for stat in base_metadata.statistics):
         raise ValueError(f"Statistics with snapshot id {update.snapshot_id} does not exist")
 
-    statistics = reject_statistics(base_metadata.statistics, update.snapshot_id)
+    statistics = filter_statistics_by_snapshot_id(base_metadata.statistics, update.snapshot_id)
     context.add_update(update)
 
     return base_metadata.model_copy(update={"statistics": statistics})