Skip to content

HIVE-29529: Iceberg: Load table from cache HiveIcebergOutputCommitter #6394

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
ayushtkn wants to merge 2 commits into
base: master
Choose a base branch
from
+3 −10
Open

HIVE-29529: Iceberg: Load table from cache HiveIcebergOutputCommitter #6394

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
99a9d99
HIVE-29529: Iceberg: Load table from cache HiveIcebergOutputCommitter
ayushtkn Mar 30, 2026
7528230
Refactor getSnapshoId()
ayushtkn Mar 30, 2026
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
13 changes: 3 additions & 10 deletions .../iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergOutputCommitter.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -440,7 +440,7 @@ private void commitTable(FileIO io, ExecutorService executor, OutputTable output
for (JobContext jobContext : jobContexts) {
JobConf conf = jobContext.getJobConf();

table = Optional.ofNullable(table).orElseGet(() -> Catalogs.loadTable(conf, catalogProperties));
table = Optional.ofNullable(table).orElseGet(() -> IcebergTableUtil.getTable(conf, catalogProperties));
Copy link
Copy Markdown
Member

@deniskuzZ deniskuzZ Mar 30, 2026 •
edited
Loading

Choose a reason for hiding this comment

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

i think we need a fresh copy, to check conflicts, not a cached one

Copy link
Copy Markdown
Member Author

@ayushtkn ayushtkn Mar 30, 2026 •
edited
Loading

Choose a reason for hiding this comment

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

should it not break at time of commit? if our commit is conflicting? Means this is again reloading the table & reading the metadata which we already read once.

Between this and before we commit there could be another commit which could create conflict. I don't think we are under lock

In that case I believe some test would fail....

Copy link
Copy Markdown
Member

@deniskuzZ deniskuzZ Mar 30, 2026 •
edited
Loading

Choose a reason for hiding this comment

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

let's see, i think it was intentional. there was no conflict found during commit. but if that's not needed, we could save some calls

Copy link
Copy Markdown
Member

@deniskuzZ deniskuzZ Mar 30, 2026

Choose a reason for hiding this comment

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

cc @difin didn't you have some issue with that in a past?

Copy link
Copy Markdown
Member

@deniskuzZ deniskuzZ Mar 30, 2026 •
edited
Loading

Choose a reason for hiding this comment

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

with that change, if it works

    Table snapshotTable = Optional.ofNullable(IcebergAcidUtil.getTransaction(table))
        .map(Transaction::table)
        .orElse(table);

is not longer needed and could be simplified

branchName = conf.get(InputFormatConfig.OUTPUT_TABLE_SNAPSHOT_REF);
snapshotId = getSnapshotId(outputTable.table, branchName);

Expand Down Expand Up @@ -518,15 +518,8 @@ private void commitTable(FileIO io, ExecutorService executor, OutputTable output
}

private Long getSnapshotId(Table table, String branchName) {
// Use the transaction table (if available) so that intra-txn changes
// (e.g. a prior INSERT) don't cause false conflict validation failures.
Table snapshotTable = Optional.ofNullable(IcebergAcidUtil.getTransaction(table))
.map(Transaction::table)
.orElse(table);

return Optional.ofNullable(IcebergTableUtil.getTableSnapshot(snapshotTable, branchName))
.map(Snapshot::snapshotId)
.orElse(null);
Snapshot snapshot = IcebergTableUtil.getTableSnapshot(table, branchName);
return Optional.ofNullable(snapshot).map(Snapshot::snapshotId).orElse(null);
}

/**
Expand Down
Loading