Pandas - Chech if two columns matches regex

I had parsed tables data from millins of PDFs.
As it a sensitive data, I tryed my best to show some of the particual situations of this

Sample DataFrame

New company_code checking two columns

pattern like 01.234.567/0001-05

I want to create a new columns, company_code, checking Regular expressions.

• If col1 matches the pattern then return the string that is col1
• If col2 matches the pattern then return the string that is col2
• else: return '-'

The result expected is

I had tried a lot to doing this. Checked many youtube channels, stack posts... But cannot solve.

If someone can help me.... Thanks a lot!

CodePudding user response：

You can try np.select

m1 = df['col1'].str.contains('^\d{2}\.\d{3}\.\d{3}/\d{4}-\d{2}$', na=False)
m2 = df['col2'].str.contains('^\d{2}\.\d{3}\.\d{3}/\d{4}-\d{2}$', na=False)

df['company_code'] = np.select(
    [m1, m2],
    [df['col1'], df['col2']],
    '-'
)

print(df['company_code'])

0                      -
1                      -
2     01.234.567/0001-05
3     23.456.789/0001-07
4     32.586.709/0001-01
5     57.888.999/0003-01
6     00.777.888/0001-01
7                      -
8                      -
9                      -
10                     -
11    01.234.567/0001-05
12    23.456.789/0001-07
13    32.586.709/0001-01
14    57.888.999/0003-01
15    00.777.888/0001-01
16                     -
Name: company_code, dtype: object