Python pandas with lambda apply difficulty

问题:

I am running the following function but somehow struggling to have it take the length condition into account (the if part). It simply runs the first part if the function only:

stringDataFrame.apply(lambda x: x.str.replace(r'[^0-9]', '') if (len(x) >= 7) else x)

it somehow only runs the x.str.replace(r'[^0-9]', '') part for some reason, what am I doing wrong here i have been stuck.

回答1:

You can use applymap when you need to work on each value separately, because apply works with all column (Series).

Then instead of using str.replace, use re.sub which works nicer for regexs:

print (stringDataFrame.applymap(lambda x: re.sub(r'[^0-9]', '', x) if (len(x) >= 7) else x))

Sample:

import pandas as pd
import re

stringDataFrame = pd.DataFrame({'A':['gdgdg454dgd','147ooo2', '123ss45678'],
                                'B':['gdgdg454dgd','x142', '12345678a'],
                                'C':['gdgdg454dgd','xx142', '12567dd8']})

print (stringDataFrame)
             A            B            C
0  gdgdg454dgd  gdgdg454dgd  gdgdg454dgd
1      147ooo2         x142        xx142
2   123ss45678    12345678a     12567dd8

print (stringDataFrame.applymap(lambda x: re.sub(r'[^0-9]', '', x) if (len(x) >= 7) else x))
          A         B       C
0       454       454     454
1      1472      x142   xx142
2  12345678  12345678  125678