Skip to content

Commit 50d2c20

Browse files
committed
fix
1 parent 3b60138 commit 50d2c20

File tree

2 files changed

+6
-16
lines changed

2 files changed

+6
-16
lines changed

paimon-python/pypaimon/read/scanner/full_starting_scanner.py

Lines changed: 5 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -32,8 +32,6 @@
3232
from pypaimon.read.split import Split
3333
from pypaimon.snapshot.snapshot_manager import SnapshotManager
3434
from pypaimon.table.bucket_mode import BucketMode
35-
from pypaimon.table.row.generic_row import GenericRow
36-
from pypaimon.manifest.schema.simple_stats import SimpleStats
3735

3836

3937
class FullStartingScanner(StartingScanner):
@@ -77,28 +75,26 @@ def scan(self) -> Plan:
7775
splits = self._apply_push_down_limit(splits)
7876
return Plan(splits)
7977

80-
def _read_manifest_files(self) -> List[ManifestFileMeta]:
78+
def plan_files(self) -> List[ManifestEntry]:
8179
latest_snapshot = self.snapshot_manager.get_latest_snapshot()
8280
if not latest_snapshot:
8381
return []
8482
manifest_files = self.manifest_list_manager.read_all(latest_snapshot)
85-
return self.filter_manifest_files(manifest_files)
83+
return self.read_manifest_entries(manifest_files)
8684

87-
def filter_manifest_files(self, files: List[ManifestFileMeta]) -> List[ManifestFileMeta]:
85+
def read_manifest_entries(self, manifest_files: List[ManifestFileMeta]) -> List[ManifestEntry]:
8886
def filter_manifest_file(file: ManifestFileMeta) -> bool:
8987
if not self.partition_key_predicate:
9088
return True
9189
return self.partition_key_predicate.test_by_simple_stats(
9290
file.partition_stats,
9391
file.num_added_files + file.num_deleted_files)
9492

95-
return [file for file in files if filter_manifest_file(file)]
96-
97-
def plan_files(self) -> List[ManifestEntry]:
98-
manifest_files = self._read_manifest_files()
9993
deleted_entries = set()
10094
added_entries = []
10195
for manifest_file in manifest_files:
96+
if not filter_manifest_file(manifest_file):
97+
continue
10298
manifest_entries = self.manifest_file_manager.read(
10399
manifest_file.file_name,
104100
lambda row: self._filter_manifest_entry(row))

paimon-python/pypaimon/read/scanner/incremental_starting_scanner.py

Lines changed: 1 addition & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -43,13 +43,7 @@ def plan_files(self) -> List[ManifestEntry]:
4343
for snapshot in snapshots_in_range:
4444
# Get manifest files for this snapshot
4545
manifest_files = self.manifest_list_manager.read_delta(snapshot)
46-
manifest_files = self.filter_manifest_files(manifest_files)
47-
48-
# Read all entries from manifest files
49-
for manifest_file in manifest_files:
50-
entries = self.manifest_file_manager.read(manifest_file.file_name,
51-
lambda row: self._filter_manifest_entry(row))
52-
file_entries.extend(entries)
46+
file_entries.extend(self.read_manifest_entries(manifest_files))
5347
return file_entries
5448

5549
@staticmethod

0 commit comments

Comments
 (0)