How can I convert following pandas series to floats:
s = pd.Series(['2,3', 5.1, '', '3.2'])
I tried s.to_float() but got a ValueError: Unable to parse string "2,3" at position 0
Asked
Active
Viewed 109 times
0
-
dupe: https://stackoverflow.com/questions/15891038/pandas-change-data-type-of-columns – EdChum Jun 05 '17 at 14:35
-
Do you want to treat `2,3` as `2.3`? – ASGM Jun 05 '17 at 14:36
-
indeed I want 2,3 as 2.3 – ronnydw Jun 05 '17 at 14:38
-
@EdChum nope, I have mixed formats for decimals, with ',' and '.' as decimal point – ronnydw Jun 05 '17 at 14:40
1 Answers
1
Cast the Series to str if necessary, then str.replace the comma with a decimal, then you can use to_numeric:
In[24]:
s = pd.to_numeric(s.astype(str).str.replace(',','.'))
s
Out[24]:
0 2.3
1 5.1
2 NaN
3 3.2
dtype: float64
If the dtype of the Series is already str then the astype(str) step can be skipped, but I'm assuming you have a more complicated real dataset issue
EdChum
- 376,765
- 198
- 813
- 562