Fix issue with reading footer of orc file with large stripe

[anton-kutuzov]

Description

We have tables in dds layer that look like

CREATE TABLE dwh.dds.some_table (
   id1 bigint,
   flag boolean,
   id2 bigint
)
WITH (
   external_location = 's3a://some_path',
   format = 'ORC'
)

where id1 is unique and id2 has only two values. The data file has 70 MB size and more than 7 000 000 000 rows.
When we try to read data the error occurs

TrinoExternalError(type=EXTERNAL, name=HIVE_CANNOT_OPEN_SPLIT, message="Error opening Hive split s3a://some_path/some_table.orc (offset=0, length=71782995): integer overflow", query_id= 20250417_090308_49505_msr65)

It happens because OrcProto support long type, but in the trino code we have int type.

Release notes

( ) This is not user-visible or is docs only, and no release notes are required.
( ) Release notes are required. Please propose a release note for me.
(x) Release notes are required, with the following suggested text:

## Hive, Iceberg
* Fix query failure when reading ORC files with a large row count. ({issue}`25634`)

[cla-bot]

Thank you for your pull request and welcome to our community. We could not parse the GitHub identity of the following contributors: Anton Kutuzov.
This is most likely caused by a git client misconfiguration; please make sure to:

1. check if your git client is configured with an email to sign commits git config --list | grep email
2. If not, set it up using git config --global user.email [email protected]
3. Make sure that the git commit email is configured in your GitHub account settings, see https://github.com/settings/emails

[raunaqmorarka]

Can you show the full error stacktrace from the failed query ?
I don't think Trino would write orc files with such large stripe row count, so I'd expect the code changes to be mostly in the reader classes.

[anton-kutuzov]

Can you show the full error stacktrace from the failed query ? I don't think Trino would write orc files with such large stripe row count, so I'd expect the code changes to be mostly in the reader classes.

Yes, of course. We write this files using other tools.
There is error stacktrace:

"type": "io.trino.spi.TrinoException",
  "message": "Error opening Hive split s3a://path/file.orc (offset=0, length=53002489): integer overflow",
  "cause": {
    "type": "java.lang.ArithmeticException",
    "message": "integer overflow",
    "suppressed": [],
    "stack": [
      "java.base/java.lang.Math.toIntExact(Math.java:1372)",
      "io.trino.orc.metadata.OrcMetadataReader.toStripeInformation(OrcMetadataReader.java:166)",
      "java.base/java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:215)",
      "java.base/java.util.ArrayList$ArrayListSpliterator.forEachRemaining(ArrayList.java:1709)",
      "java.base/java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:570)",
      "java.base/java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:560)",
      "java.base/java.util.stream.ReduceOps$ReduceOp.evaluateSequential(ReduceOps.java:921)",
      "java.base/java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:265)",
      "java.base/java.util.stream.ReferencePipeline.collect(ReferencePipeline.java:727)",
      "io.trino.orc.metadata.OrcMetadataReader.toStripeInformation(OrcMetadataReader.java:160)",
      "io.trino.orc.metadata.OrcMetadataReader.readFooter(OrcMetadataReader.java:149)",
      "io.trino.orc.metadata.ExceptionWrappingMetadataReader.readFooter(ExceptionWrappingMetadataReader.java:74)",
      "io.trino.orc.OrcReader.<init>(OrcReader.java:205)",
      "io.trino.orc.OrcReader.createOrcReader(OrcReader.java:118)",
      "io.trino.orc.OrcReader.createOrcReader(OrcReader.java:95)",
      "io.trino.plugin.hive.orc.OrcPageSourceFactory.createOrcPageSource(OrcPageSourceFactory.java:269)",
      "io.trino.plugin.hive.orc.OrcPageSourceFactory.createPageSource(OrcPageSourceFactory.java:194)",
      "io.trino.plugin.hive.HivePageSourceProvider.createHivePageSource(HivePageSourceProvider.java:203)",
      "io.trino.plugin.hive.HivePageSourceProvider.createPageSource(HivePageSourceProvider.java:138)",
      "io.trino.plugin.base.classloader.ClassLoaderSafeConnectorPageSourceProvider.createPageSource(ClassLoaderSafeConnectorPageSourceProvider.java:48)",
      "io.trino.split.PageSourceManager$PageSourceProviderInstance.createPageSource(PageSourceManager.java:79)",
      "io.trino.operator.TableScanOperator.getOutput(TableScanOperator.java:265)",
      "io.trino.operator.Driver.processInternal(Driver.java:403)",
      "io.trino.operator.Driver.lambda$process$8(Driver.java:306)",
      "io.trino.operator.Driver.tryWithLock(Driver.java:709)",
      "io.trino.operator.Driver.process(Driver.java:298)",
      "io.trino.operator.Driver.processForDuration(Driver.java:269)",
      "io.trino.execution.SqlTaskExecution$DriverSplitRunner.processFor(SqlTaskExecution.java:890)",
      "io.trino.execution.executor.timesharing.PrioritizedSplitRunner.process(PrioritizedSplitRunner.java:187)",
      "io.trino.execution.executor.timesharing.TimeSharingTaskExecutor$TaskRunner.run(TimeSharingTaskExecutor.java:565)",
      "io.trino.$gen.Trino_461____20250416_210446_2.run(Unknown Source)",
      "java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144)",
      "java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:642)",
      "java.base/java.lang.Thread.run(Thread.java:1575)"
    ],
    "errorCode": {
      "code": 65536,
      "name": "GENERIC_INTERNAL_ERROR",
      "type": "INTERNAL_ERROR",
      "fatal": false
    }
  }

The code changes in writer classes only in places where the numberOfRows used, because the type was changed from int to long as in OrcProto.

[anton-kutuzov]

@raunaqmorarka could you please merge PR? Or do you wait something from me?

[Copilot]

Pull Request Overview

This PR fixes an issue with reading ORC files that have a very large stripe by updating the data types and method signatures from int to long where necessary to prevent integer overflow.

• Update test coverage to verify large stripe row counts
• Change field types and method parameters from int to long
• Adjust calculations (e.g. ceil function) accordingly in multiple modules

Reviewed Changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
lib/trino-orc/src/test/java/io/trino/orc/metadata/TestOrcMetadataReader.java	Adds a test verifying correct handling of large row counts
lib/trino-orc/src/main/java/io/trino/orc/metadata/StripeInformation.java	Updates the numberOfRows field and related accessors from int to long
lib/trino-orc/src/main/java/io/trino/orc/metadata/OrcMetadataReader.java	Adapts conversion to use long values for stripe row counts
lib/trino-orc/src/main/java/io/trino/orc/StripeReader.java	Changes parameters and local variable types from int to long and revises the ceil function
lib/trino-orc/src/main/java/io/trino/orc/OrcWriterStats.java	Updates the signature for stripeRows to long
lib/trino-orc/src/main/java/io/trino/orc/OrcWriterFlushStats.java	Updates the signature for stripeRows to long
lib/trino-orc/src/main/java/io/trino/orc/OrcWriteValidation.java	Revises addStripe method to accept long values
lib/trino-orc/src/main/java/io/trino/orc/OrcRecordReader.java	Updates helper method validateWriteStripe to use long