Fritz Larco

Hey @colinmollenhour, thanks for reporting. Unfortunately, it's a bit difficult to pin it down without the error stack details. If possible, can you run with the env vars (using ver...

@colinmollenhour I suspect this issue should be solved by https://github.com/slingdata-io/sling-cli/issues/668 There was an issue with the buffer size when processing large one-liner blobs. This is fixed for next release. I'm...

hey @noat28 i'm wondering if something is happening in the duckdb process while reading those parquet files. Can you answer the following questions: 1. Does it get stuck deterministically? As...

@noat28 thanks, but in your script you're reading with pandas. Sling uses [DuckDB CLI](https://duckdb.org/install/) to read your parquet files. Could you try to read from duckdb so it uses the...

thanks @noat28 , that didn't help me pin down the issue. Could you try with the latest [dev build](https://github.com/slingdata-io/sling-cli?tab=readme-ov-file#automated-dev-builds)? There is new function to dump the stack of the program...

@AndyCokeZero thanks, super-helpful. I'll try to reproduce.

@AndyCokeZero I can't reproduce the hang-up. I have a python file using pyarrow to generate the large values (over 66K chars). it just works as normal for me... Do you...

@AndyCokeZero thank you again. Was able to reproduce. It all came down to the buffer size for **one line**, when duckDB CLI sends to Sling CLI. Buffer size was 64kb,...

Interesting, that should work since cell values including a comma would be enclosed with a `"`, can you try `format: json` in your env.yaml starrocks connection ? It'll be slower,...

I just added `KEEP_TEMP_FILES` here: https://github.com/slingdata-io/sling-cli/pull/344/commits/65a6119b09ad15e84f387d9f3ea028cb6b97b4be If you build the binary on that branch, and pass env var `KEEP_TEMP_FILES=true`, it'll keep it (this is already available in other connections like...