Iceberg, support for sorting

Iceberg 1.x and quite some query engine already support sorting on writing/optimizing parquet files. Dremio does not. As sorting is a very essential instrument to optimize file size, minimize reads and finally boost query performance, I wonder why Dremio - always hunting for best in class performance - is ignoring this very simple but effective capability.

What is your proposal or recommended work around to overcome this missing capability? We want and need sort our parquet files when optimizing them.

PS.: Microsoft‘s proprietary V-Order for sorting parquet files on write is said to yield a 2x faster on average query execution.

It is in the works this very moment.

1 Like

I should have known :wink: Many thanks.

I assume it will be somehow announced once implemented.