INT96 timestamp is read as OffsetDateTime, not LocalDateTime

[anesterenok]

Apache Iceberg version

1.8.0 (latest release)

Query engine

Other

Please describe the bug 🐞

When a Parquet int64 timestamp is read, it is read as LocalDateTime (correct):

iceberg/parquet/src/main/java/org/apache/iceberg/data/parquet/GenericParquetReaders.java

Line 84 in c277c20

static class TimestampReader extends ParquetValueReaders.PrimitiveReader<LocalDateTime> {

However, when an int96 timestamp is read, it is read as OffsetDateTime (wrong):

iceberg/parquet/src/main/java/org/apache/iceberg/data/parquet/GenericParquetReaders.java

Line 106 in c277c20

static class TimestampInt96Reader extends ParquetValueReaders.PrimitiveReader<OffsetDateTime> {

This is wrong because OffsetDateTime is a type for timestamps with time zone, like (correct):

iceberg/parquet/src/main/java/org/apache/iceberg/data/parquet/GenericParquetReaders.java

Line 126 in c277c20

static class TimestamptzReader extends ParquetValueReaders.PrimitiveReader<OffsetDateTime> {

Expected: all timestamp reads return LocalDateTime, all timestamps with time zone reads return OffsetDateTime. Because it is the schema that defines how a value is interpreted, not its storage format.

As for fixing this: if it's just changing GenericParquetReaders (and a test), I can do a PR. If it's something more substantial, I'm afraid I cannot help.

Willingness to contribute

• I can contribute a fix for this bug independently
• I would be willing to contribute a fix for this bug with guidance from the Iceberg community
• I cannot contribute a fix for this bug at this time