Improve TypeError message for str.cat

[h-vetinari]

Currently,

s = pd.Series(['a', 'b', 'c'])
s.str.cat([1, 2, 3])

yields

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\Axel Obermeier\eclipse-workspace\pddev\pandas\core\strings.py", line 2222, in cat
    res = str_cat(data, others=others, sep=sep, na_rep=na_rep)
  File "C:\Users\Axel Obermeier\eclipse-workspace\pddev\pandas\core\strings.py", line 111, in str_cat
    cats = [sep.join(tup) for tup in tuples]
  File "C:\Users\Axel Obermeier\eclipse-workspace\pddev\pandas\core\strings.py", line 111, in <listcomp>
    cats = [sep.join(tup) for tup in tuples]
TypeError: sequence item 1: expected str instance, int found

IMO, this should be improved to have a better error message, and shallower stack trace.

[WillAyd]

What are you suggesting exactly? Error message reflects what you would get from a standard Python operation:

>>> "".join(['foo', 1])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: sequence item 1: expected str instance, int found

[h-vetinari]

@WillAyd

That's of course correct, but users are not dealing with str/int instances, but usually with Series/ndarray, etc., and so the mentioned warning gives the right hint, but is not necessarily the best we can do. I opened this issue as it was a side product of cleaning up some internals/tests for #22725.

[h-vetinari]

@jreback

I'll open another PR for this, but this will not be closed by #22725 (was split off due to review)

[h-vetinari]

@WillAyd The code for this was splitt off of #22725 to focus that PR (I had edited the OP to reflect that) - this issue is still open.