As mentioned in a discussion for pandas, there is work going on to reduce the package size of pyarrow (e.g. to split it up in a package with core functionality), compare pandas-dev/pandas#54466 (comment)
If this has happened, we should switch to using this smaller package as a dependency of audb.
As mentioned in a discussion for
pandas, there is work going on to reduce the package size ofpyarrow(e.g. to split it up in a package with core functionality), compare pandas-dev/pandas#54466 (comment)If this has happened, we should switch to using this smaller package as a dependency of
audb.