How to split one column with mixed values of integers and strings into two different columns using p-CodePudding

HEADER A	header B
test	A
test1	B
test2	1
test3	E
test4	2

INTO

HEADER A	header c	header d
test	A	Nan
test1	B	Nan
test2	Nan	1
test3	E	Nan
test4	Nan	2

In short i want to split header B INTO header c and header d as shown in the table. Using pandas

CodePudding user response：

You can use pandas.Series.str.isnumeric then with Series.mask

df['header c'] = df['header B'].mask(df['header B'].str.isnumeric(), np.nan)
df['header d'] = df['header B'].where(df['header B'].str.isnumeric(), np.nan)

print(df)

  HEADER A header B header c header d
0     test        A        A      NaN
1    test1        B        B      NaN
2    test2        1      NaN        1
3    test3        E        E      NaN
4    test4        2      NaN        2

CodePudding user response：

If you have strings, use str.extract:

df[['header c', 'header b']] = df['header B'].str.extract(r'^(\D )|(\d )$')

output:

  HEADER A  header B header c header b
0     test         A        A      NaN
1    test1         B        B      NaN
2    test2         1      NaN        1
3    test3         E        E      NaN
4    test4         2      NaN        2