Iceberg configuration properties 'iceberg.compression-codec description is inaccurate

[PingLiuPing]

https://prestodb.github.io/docs/0.294/connector/iceberg.html#configuration-properties

The property iceberg.compression-code says it supports NONE, SNAPPY, GZIP, LZ4, and ZSTD. And also the default file format iceberg.file-format is PARQUET.

But from the code I see ZSTD and LZ4 are not supported for parquet format.

https://github.com/prestodb/presto/blob/master/presto-hive/src/main/java/com/facebook/presto/hive/HiveCompressionCodec.java#L35-L36

presto:iceberg> CREATE TABLE customer ( custkey BIGINT NOT NULL, name VARCHAR(25) NOT NULL, address VARCHAR(40) NOT NULL, nationkey BIGINT NOT NULL, phone VARCHAR(15) NOT NULL, acctbal DECIMAL(12,2) NOT NULL, mktsegment VARCHAR(10) NOT NULL, comment VARCHAR(117) NOT NULL ) WITH (format = 'PARQUET');
Query 20251009_134232_00012_jxfcc failed: Compression codec ZSTD is not supported for Parquet format

Your Environment

• Presto version used:
• Storage (HDFS/S3/GCS..):
• Data source and connector used:
• Deployment (Cloud or On-prem):
• Pastebin link to the complete debug logs:

Expected Behavior

Current Behavior

Possible Solution

Steps to Reproduce

Screenshots (if appropriate)

Context

[steveburnett]

As the description for iceberg.file-format states "The available values are PARQUET and ORC.", would adding the following to the description for iceberg.compression-codec address this concern? (assuming the following statement is true)

"Note: ZSTD and LZ4 are only available if iceberg.file-format=ORC."

[PingLiuPing]

@steveburnett I think so. We need a note to indicate which compression codec are supported for each file format.

[hantangwangd]

In PR #25021, @ZacBlanco attempted to add support for ZSTD and LZ4 on parquet. It turned out that ZSTD on parquet can be supported, while LZ4 on parquet still has some underlying problems to resolve.

[PingLiuPing]

@hantangwangd Thanks for the info. I think we can update the doc first and once #25021 been merged we update the doc again.What's your opinion?

[hantangwangd]

@PingLiuPing Sounds good to me!

[PingLiuPing]

@hantangwangd Not sure if you know any history of why the default compression codec is GZIP for parquet?

[hantangwangd]

@PingLiuPing It seems that Iceberg use GZIP as the default compression for parquet before version1.4. See here.

[hantangwangd]

When iceberg connector was first added by @ChunxuTang, the version of Iceberg was 0.9.0.