Skip to content

[improvement](fe) Skip zero-length files in FileSplitter to avoid sending empty splits to BE #62482

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
kaka11chen wants to merge 1 commit into
base: master
Choose a base branch
from
+28 −0
Open

[improvement](fe) Skip zero-length files in FileSplitter to avoid sending empty splits to BE #62482

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions fe/fe-core/src/main/java/org/apache/doris/datasource/FileSplitter.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,6 +101,10 @@ public List<Split> splitFile(
List<String> partitionValues,
SplitCreator splitCreator)
throws IOException {
if (length <= 0) {
// Zero-length files contain no data; skip to avoid sending empty splits to BE.
return Lists.newArrayList();
Copy link

Copilot AI Apr 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The early return for zero-length files currently allocates a new mutable list on every call. Since this is a hot path when scanning many files, consider returning Collections.emptyList() (or ImmutableList.of()) to avoid unnecessary allocations and signal immutability.

Suggested change
return Lists.newArrayList();
return ImmutableList.of();

Copilot uses AI. Check for mistakes.
}
// Pass splitCreator.create() to set target file split size to calculate split weight.
long targetFileSplitSize = specifiedFileSplitSize > 0 ? specifiedFileSplitSize : maxSplitSize;
if (blockLocations == null) {
Expand Down
24 changes: 24 additions & 0 deletions fe/fe-core/src/test/java/org/apache/doris/datasource/FileSplitterTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -195,6 +195,30 @@ public void testNullRemainingInitialSplitIsAllowed() throws Exception {
Assert.assertEquals(1, splits.size());
}

@Test
public void testZeroLengthFileProducesNoSplits() throws Exception {
LocationPath loc = LocationPath.of("hdfs://example.com/path/emptyfile");
BlockLocation[] locations = new BlockLocation[]{new BlockLocation(null, new String[]{"h1"}, 0L, 0L)};
FileSplitter fileSplitter = new FileSplitter(32 * MB, 64 * MB, DEFAULT_INITIAL_SPLITS);
// Non-splittable zero-length file
List<Split> splits = fileSplitter.splitFile(
loc, 0L, locations, 0L, 0L, false,
Collections.emptyList(), FileSplit.FileSplitCreator.DEFAULT);
Assert.assertTrue("Zero-length file should produce no splits", splits.isEmpty());
// Splittable zero-length file
splits = fileSplitter.splitFile(
loc, 0L, locations, 0L, 0L, true,
Collections.emptyList(), FileSplit.FileSplitCreator.DEFAULT);
Assert.assertTrue("Zero-length splittable file should produce no splits", splits.isEmpty());
// Null block locations with zero-length file
splits = fileSplitter.splitFile(
loc, 0L, null, 0L, 0L, true,
Collections.emptyList(), FileSplit.FileSplitCreator.DEFAULT);
Assert.assertTrue("Zero-length file with null locations should produce no splits", splits.isEmpty());
// Counter should not be decremented for skipped zero-length files
Assert.assertEquals(DEFAULT_INITIAL_SPLITS, fileSplitter.getRemainingInitialSplitNum());
}

@Test
public void testSmallFileNoSplit() throws Exception {
LocationPath loc = LocationPath.of("hdfs://example.com/path/small");
Expand Down
Loading