Alexander Nenninger

icon company NTT DATA icon location München icon location Still trying to figure out how all of this works.

Results 4 comments of Alexander Nenninger

Add BytesIO support to `scan_csv`

+1 to this. Implementing this proposal would allow for scanning memory-mapped files that need additional processing, e.g. parsing custom headers.

Allow Index Column of Type Duration in DataFrame.rolling

What the heck? `Codecov token not found`??: https://github.com/pola-rs/polars/actions/runs/8913916408/job/24480392846?pr=15999

Allow Index Column of Type Duration in DataFrame.rolling

Guess this won’t happen then.

Scanning, Filtering and Sinking Parquet with `PartitionMaxSize` Causes Excessive Memory Use

@nameexhaustion With materialization in memory: ``` Line # Mem usage Increment Occurrences Line Contents ============================================================= 49 669.9 MiB 669.9 MiB 1 @profile 50 def memory(input_dir): 51 9665.3 MiB 0.3 MiB...