ENH: Support reading from S3

[pjob]

Tests not written. Will need a way to mock boto.

[ghost]

IMO this doesn't belong in pandas (neither are other API storage services),
users can implement that in their own scripts.

[pjob]

Thanks and I understand your point.

Retrieving files from S3 offers a powerful io experience in pandas given the way people use S3 for logs and other historical record and pandas' handling of compressed files. Perhaps this would be better as its own IO tool similar to the sql wrappers. It would aid the user in retrieving from and setting to S3.

Feel free to close this or move the discussion to a better resource (mailing list, issue tracker).

[ghost]

This has me really stumped for some reason. But I guess I should err on the side of "it's useful",
and we do want to make it as easy to get data into pandas as possible.

Please add a note in RELEASE.rst, and mention boto as an optional dependency in README.md
for this functionality, I'll merge this for 0.11.1.

[nehalecky]

That's the spirit all!

[pjob]

@y-p Thank you. Everything should be up to date.

[ghost]

@pjob , can you please add a blurb to io.rst in the docs?

[pjob]

@y-p sure.

[ghost]

Thanks.