Shekharrajak opened a new issue, #60789:
URL: https://github.com/apache/doris/issues/60789
Doris currently uses a **hybrid Java FE + custom C++ BE** architecture for
Iceberg:
- **FE (Java):** Uses the official `iceberg-java` library for catalog
access, scan planning,
snapshot management, and metadata operations. Produces `TFileRangeDesc`
scan ranges via Thrift.
- **BE (C++):** Contains a hand-rolled, minimal Iceberg spec
implementation under
`doris::iceberg` (~14 C++ files) covering types, schema, partition spec
parsing,
and Arrow schema conversion. The read path (`IcebergTableReader`)
delegates to Doris's
own Parquet/ORC readers with Iceberg-specific handling for position
deletes, equality
deletes, and deletion vectors. The write path (`VIcebergTableWriter`)
handles partitioned
writes with transform support.
Meanwhile, [apache/iceberg-cpp](https://github.com/apache/iceberg-cpp)
(v0.3.0-SNAPSHOT) is
a growing, spec-compliant C++ implementation with significantly broader
coverage.
Adopt `apache/iceberg-cpp` as a third-party dependency in Doris BE to
progressively replace
the custom `doris::iceberg` namespace.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
