lineterminator in to_csv doesn't work

[cpcloud]

In [3]: df = pd.DataFrame({'a': ['3.4', '34.400'], 'b': ['5,5', '10,200']})

In [4]: df
Out[4]:
        a       b
0     3.4     5,5
1  34.400  10,200

In [5]: df.to_csv(lineterminator='-')
Out[5]: ',a,b\n0,3.4,"5,5"\n1,34.400,"10,200"\n'

I would've expected

Out[5]: ',a,b-0,3.4,"5.5"-1,34.400,"10,200"-'

[onesandzeroes]

The argument in DataFrame.to_csv() is actually line_terminator, not lineterminator. A little confusing, since csv.writer uses the latter. The argument gets changed here.

Then I guess that confusion gets compounded by the fact that DataFrame.to_csv() accepts extra keyword args with a **kwds argument, but then doesn't seem to validate them. It only seems to use engine and cols from **kwds, so maybe the real issue here is that it should be warning if given unrecognized args?

[mangecoeur]

definitely bad behaviour to accept invalid kwargs! if only engine and cols are used they should just be explicitly defined.

[gfyoung]

@jreback : the kwds parameter is not used anymore in to_csv - can't we just remove it now (0.19.0)?

[jorisvandenbossche]

@gfyoung yep!