Pandas Strip numbers from a string column in python

Question

I have a column in pandas which has string and numbers mixed I want to strip numbers from the string.

A
11286011
11268163
C7DDA72897
C8ABC557
Abul
C80DAS577
C80DSS665

Want an output as

A
C7DDA72897
C8ABC557
Abul
C80DAS577
C80DSS665

Possible duplicate of [remove non-numeric rows in one column with pandas](http://stackoverflow.com/questions/33961028/remove-non-numeric-rows-in-one-column-with-pandas) — Andy, Apr 13 '17 at 13:26
Possible duplicate of [Pandas select only numeric or integer field from dataframe](http://stackoverflow.com/questions/24883503/pandas-select-only-numeric-or-integer-field-from-dataframe) — dot.Py, Apr 13 '17 at 13:27

score 4 · Accepted Answer · answered Apr 13 '17 at 13:27

4

In [52]: df
Out[52]:
            A
0    11286011
1    11268163
2  C7DDA72897
3    C8ABC557
4   C80DAS577
5   C80DSS665

In [53]: df = pd.to_numeric(df.A, errors='coerce').dropna()

In [54]: df
Out[54]:
0    11286011.0
1    11268163.0
Name: A, dtype: float64

or using RegEx:

In [59]: df.loc[~df.A.str.contains(r'\D+')]
Out[59]:
          A
0  11286011
1  11268163

answered Apr 13 '17 at 13:27

MaxU - stop genocide of UA

191,778
30
340
375

Thanks @MaxU your RegEx solved the problem for me. Works great. – Abul Apr 19 '17 at 05:26

piRSquared · Answer 2 · 2017-04-13T13:33:46.990

4

You can use .str.isnumeric to use in boolean slicing.

df[df.A.astype(str).str.isnumeric()]

          A
0  11286011
1  11268163

As pointed out by @MaxU, assuming every element is already a string, you can limit this to

df[df.A.str.isnumeric()]

          A
0  11286011
1  11268163

edited Apr 13 '17 at 13:33

answered Apr 13 '17 at 13:28

piRSquared

265,629
48
427
571

I think `A` is already of string (object) dtype, so we can omit `.astype(str)`... – MaxU - stop genocide of UA Apr 13 '17 at 13:31
1

@MaxU `dtype==object` doesn't imply that all elements are strings. You can have an actual float or int in the array. If you do, `str.isnumeric` will fail. – piRSquared Apr 13 '17 at 13:51
this is a good point! I just have tested it - `.isnumeric()` will retyren `NaN` in this case – MaxU - stop genocide of UA Apr 13 '17 at 14:04
@piRSquared Thanks for the help too. – Abul Apr 19 '17 at 10:40

Pandas Strip numbers from a string column in python

2 Answers2