DOC GH3508 (bis) added basic documentation of google analytics in remot...

[benoitpointet]

added the GA doc to the correct file: "remote_data.rst"
closes #3508

[jreback]

actually if you would add a release note (in Enhancements) in v0.15.2, with a pointer to the docs (e.g. see here)....then people will know about the new docs.

thanks

[benoitpointet]

Added entry in whatsnew for v15.0.2.

[jorisvandenbossche]

For the rest, looks good to me!
Can you squash both commits into one?

[jreback]

merged via 9dd675b

thanks!

[jreback]

@benoitpointet can you run this PR

#8785

and see if it removes the warnings that were previously generated.

lmk

[benoitpointet]

Tested, it does indeed. It however does not remove the index_col bug described in #8773.

[jreback]

k thanks...

#8773 is not a bug, or did this use to work? (e.g. has the format changed)? The user needs to specify the index_col?

[benoitpointet]

it used to work without​ having to provide the index_col argument.
b.

[jreback]

if you want to somehow infer that the index_col is 0 that would be great!

(e.g if index_col=None is the default), then just try it, if it fails, then you can try with index_col=0). IF its unambiguous.

Ideally would have tests for this.....

[benoitpointet]

I'm unsure about hacking the read_ga function to infer such things.

If you look at the behaviour of the index_col param of the read_csv function for example, it support values such as index_col=['start', 'end'] or index_col=1. That for me should be the default behaviour of the index_col param in read_ga. I'm still having a hard time figuring out why it does not behave so ...