Oleks V

Narrowed down the problem to query ``` with t1 as (select 1 a, 2 b) select * from t1 where exists (select 1 from t1 t2 where t2.a = t1.a...

The problem is in LeftSemi/LeftAnti join types with extra join filter. So the join side gets built correctly but join filter for notEq case gets crashed

Full test to reproduce ``` #[tokio::test] async fn test_() -> Result { let ctx: SessionContext = SessionContext::new(); let sql = "set datafusion.optimizer.prefer_hash_join = false;"; let _ = ctx.sql(sql).await?.collect().await?; let sql...

Thanks @lkt for your contribution, does this PR stand for `array_agg` performance improvement? If so it is good to see a bench mark results

sounds good, what will happen if the number is without floating point? lets say I have an array `[10f64, 10.1f64]` what is the expected output?

Thanks @akubera that looks great, looking forward for the release

Likely there is a support on Seek::StartFrom on a file level ``` let decoder = if is_whole_file_scanned { // Don't seek if no range as breaks FIFO files file_compression_type.convert_read(file)? }...

Thanks @Tangruilin I think we need to find out how to deal with parallel reads and multiple files prior to development

I think the idea of skipping N rows on the file level doesn't make much sense. What we can probably do is to skip N rows on dataframe level, but...

I think you can skip lines as I posted the code above https://github.com/apache/arrow-datafusion/issues/8824#issuecomment-1907167127 But my concern is for reading multiple files, is it expected to skip lines for all the...