Jeff Reback

no that defeats the purpose of iso format detection you could allow infer_datetime_format='strict' for maybe exact='striict') which when s format is specified could bypass the inference check (and then would...

if errors=coerce then this wouldn't raise

So we ONLY hit the dateutil path (and hence use `dayfirst/yearfirst` settings) IF we fail to parse ISO8601 (IOW an exception is raised). So easiest thing to do is track...

you would do this in array_to_datetime once a date is parsed with dayfirst or yearfirst then it would be an error to not parse others in the same way

@garfieldthecat pls have a read thru the very extensive documentation this having mixed dayfirst datetimes is only a problem at parse time; you can specify multiple options to properly parse...

well see:http://pandas.pydata.org/pandas-docs/stable/user_guide/io.html#io-read-csv-table for date parsing, the doc-string: http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.to_datetime.html?highlight=to_datetime#pandas.to_datetime I am not sure what else you are refering, ``array_to_datetime`` is an internal implementation, so if you want to fix this issue,...

you are welcome to submit a PR: http://pandas.pydata.org/pandas-docs/stable/development/contributing.html

cc @anmyachev who recently update the internal parsing of dayfirst to be in cython. A patch for this issue would now be somewhat performant.

pandas is of course a completely volunteer project with almost 3000 issues, so statements like this > this is a very serious problem are not very helpful to anyhow. pull-requests...

@matthew-brett didn't really mean to call you out :-D, more specifically about prioritization of *anything* in pandas is subject to resource availability, meaning you can't prioritize, no matter how 'important'...