Add a "filename" column option when reading multiple CSVs with globbing

[Tiv0w]

Problem description

I'd like to use Polars (in Python) to read multiple CSVs at once, using a glob pl.read_csv('*.csv'), but all those files have exactly the same structure, and are only identifiable via their respective filename.

For now, my code is:

    files = glob.glob('./*.csv')
    data = pl.concat(
        [
            pl.read_csv(f).with_columns(pl.lit(os.path.basename(f)).alias("Symbol"))
            for f in files
        ]
    )

It would be great if an add_filename option could be added to the read_csv/scan_csv functions.
If set, this would include an extra filename column in the resulting DataFrame/LazyFrame, so further processing can be done with this information.

[ritchie46]

Your code is already a fine solution isn't it? No need to put it in polars itself.

[Tiv0w]

Yeah, realistically, it's fine. Thanks.