API: pandas.read_table should handle Inf the same way as inf

[aebrahim]

entries with "Inf" (beginning with upper case I) will assign that column a dtype of object while "inf" is parsed as infinity.

I believe that in R, the function write.table uses "Inf" (with the uppercase) which would make importing tables exported from R more difficult.

[jreback]

pretty straightforward to amend this work

[jtratner]

@jreback I'll have a PR soon. Pretty minimal change, esp if just special case it. Much slower if pass a list and then use the in operator.

[jreback]

no this is just a really small change in src/parser.pyx

it string matches inf and -inf using strcmp

just need to do this ignoring case

[jreback]

could prob just change strcmp to strcasecmp

and add a test