GH-41121: [C++] Fix: left anti join filter empty rows.

[Light-City]

Rationale for this change

Since the left anti filter implementation is based on the left semi filter, and an assertion error occurs when the left semi filter rows are empty, this problem should be fixed.

What changes are included in this PR?

swiss_join.cc and hash_join_node_test.cc

Are these changes tested?

Yes

Are there any user-facing changes?

No

• GitHub Issue: [C++] SwissJoin Filter Empty Rows Failed! #41121

[github-actions]

⚠️ GitHub issue #41121 has been automatically assigned in GitHub to PR creator.

[Light-City]

I don’t know why lint ci hangs without any prompt?

[ZhangHuiGui]

I don’t know why lint ci hangs without any prompt?

Seems you need to format your codes:

 TEST(HashJoin, FilterEmptyRows) {
   BatchesWithSchema input_left;
-  input_left.batches = {ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")};
+  input_left.batches = {
+      ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")};

[Light-City]

I don’t know why lint ci hangs without any prompt?

Seems you need to format your codes:

 TEST(HashJoin, FilterEmptyRows) {
   BatchesWithSchema input_left;
-  input_left.batches = {ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")};
+  input_left.batches = {
+      ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")};

thanks, I changed.

[Light-City]

test_table.py failed .The test failure has nothing to do with my PR, it seems to be asof join

[zanmato1984]

This is a great catch, thank you!

One minor suggestion though.

[zanmato1984]

test_table.py failed .The test failure has nothing to do with my PR, it seems to be asof join

Yeah, seems unrelated to this PR. Could be the same as #41149.

[zanmato1984]

+1

[zanmato1984]

We should wait for some committer to approve and merge. Thank you again for the fix!

cc @westonpace

[Light-City]

We should wait for some committer to approve and merge. Thank you again for the fix!

cc @westonpace

Okay, thank you

[mapleFU]

@zanmato1984 should this better in 16.0?

[zanmato1984]

@zanmato1984 should this better in 16.0?

Yes, it would be nice to have it in 16.0. The condition to trigger this bug is not very rare: left-semi join with no rows matched by key meanwhile filter referring left table columns only.

cc @raulcd . If we have a chance, please consider including this PR in 16.0. Thank you.

[mapleFU]

I've change this patch to blocker

[pitrou]

While this would indeed be nice to have, I've removed the "Blocker" label since this doesn't seem to be a regression.

[pitrou]

Rebased on git main for CI.

[pitrou]

@github-actions crossbow submit -g cpp

[pitrou]

This seems ok on the face of it, but I can't vouch for the correctness.

[github-actions]

Revision: 5d59164

Submitted crossbow builds: ursacomputing/crossbow @ actions-63062f6aac

Task	Status
test-alpine-linux-cpp
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-cuda-cpp
test-debian-12-cpp-amd64
test-debian-12-cpp-i386
test-fedora-39-cpp
test-ubuntu-20.04-cpp
test-ubuntu-20.04-cpp-bundled
test-ubuntu-20.04-cpp-minimal-with-formats
test-ubuntu-20.04-cpp-thread-sanitizer
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-20
test-ubuntu-22.04-cpp-emscripten
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-gcc-14

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 7 benchmarking runs that have been run so far on merge-commit 924fbec.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 7 possible false positives for unstable benchmarks that are known to sometimes produce them.