Ignore errors in pandas astype

Question

I have a numeric column that could contain another characters different form [0-9]. Say: x = pandas.Series([\"1\",\"1.2\", \"*\", \"1\", \"**.\"])

Answer 1

I think you want to use pd.to_numeric(x, errors='coerce') instead:

In [73]: x = pd.to_numeric(x, errors='coerce')

In [74]: x
Out[74]:
0    1.0
1    1.2
2    NaN
3    1.0
4    NaN
dtype: float64

PS actually x.astype(dtype = float, errors = 'ignore') - works as expected, it doesn't give an error, it just leaves series as it is as it can't convert some elements:

In [77]: x.astype(dtype = float, errors = 'ignore')
Out[77]:
0      1
1    1.2
2      *
3      1
4    **.
dtype: object   # <----- NOTE!!!

In [81]: x.astype(dtype = float, errors = 'ignore').tolist()
Out[81]: ['1', '1.2', '*', '1', '**.']