emkornfield commented on pull request #8644: URL: https://github.com/apache/arrow/pull/8644#issuecomment-729163995
> I’m not sure I understand the philosophy here though; Arrow has a lot of Python code to provide Pandas compatibility but if you always have to copy IPC data to use it from Pandas then you’re not getting any advantage from using Arrow as the underlying data format? My thinking here is Arrow isn't just about the python implementation. There are all sorts of tweaks that might make one language/library pair more performant, but changing the spec and ensuring all the other implementations can handle it is is a bunch of work, and over time degrades compatibility between implementation (we are already behind on some other features that are already in the spec). So in these cases, I tend to be conservative with changing the specification. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
