column name dedupe when already deduped returns "column_name.1.1"

[boldloop]

This isn't necessarily a bug, just something I think could be handled slightly cleaner. The code:

from io import StringIO
csv = StringIO('a,a.1,a')
pd.read_csv(csv)

returns

a	a.1	a.1.1

where I think it could return

a	a.1	a.2

I encountered this when working with data that included the columns ["Column name", "COLUMN NAME", "COLUMN NAME"]. I work in multiple steps (i.e., separating each task into a separate script that takes the output from the previous task—works for auditability), so after the first step the columns became ["Column name", "COLUMN NAME", "COLUMN NAME.1"]. Down the pipeline, I renamed all columns to use all caps for consistency, but I ended up with ["COLUMN NAME.1.1", "COLUMN NAME", "COLUMN NAME.1"]. Not a huge problem, but a little bit annoying nonetheless.

I'm going to dig in and see if I can implement this later this week in a pull request—just filing it now so that I don't forget.

[phofl]

duplicate of #14704