AlenkaF commented on issue #34994: URL: https://github.com/apache/arrow/issues/34994#issuecomment-6499446577
I did some testing myself (can reproduce) and found that the most of the cost in the first repr comes from `pa.array` importing pandas. If pandas is already imported, the first call drops from ~200 ms to ~8 ms for me locally. The issue might be in `_is_array_like` importing pandas even for plain Python list: https://github.com/apache/arrow/blob/a99464f9c625c3f85cade9d7d50eb25297ca1f8d/python/pyarrow/array.pxi#L62-L67 Claude proposes a `sys.modules` check in `_is_array_like` before importing pandas. I haven't tested or looked into this idea but think it would be a good start. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
