Iterate different Regular Expression Patterns over df Column

Question

I have a df column with text in it and I'm trying to extract different date patterns from it.

This df1 for example:

<index>    text    
0          My birthday is 10/23/89.
1          Christmas is on December 25th.
2          Thanksgiving of 11/2008 was the best.

The desired output is a 3rd column, called dates:

<index>    text                                  dates
0          My birthday is 10/23/89.               10/23/89
1          Christmas is on December 25.           25 December
2          Thanksgiving of 11/2008 was the best.  11/2008

To pull our the first date I write my first re expression, like this one:

df1['dates'] = (df1['text'].str.findall(r'\d{1,2}[/-]\d{1,2}[/-]\d{2,4}'))

And that's where I get stuck.

I don't know/understand how to write multiple re expressions and not keep writing over what's already in the df1['dates'] column.

I'd like to run my next expression:

df1['dates'] = df1['text'].str.findall(r'(?:\d{1,2})?(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)[a-z]* (?:\d{1,2}, )?\d{4}')

How, or what is the best way, to see if df['dates'] column is null and then try the next re expression?

I asked this earlier today and it got flagged as a possible duplicate of this but I think DeepSpace thought I was a lot smarter than I really am, my question is much more basic than the one he answered.

Answer 1

You can try

df['dates'] = df['text'].str.extract('.*?(\d+/\d+/?\d*).*?')


    text                                    dates
0   My birthday is 10/23/89.                10/23/89
1   Christmas is 12/25.                     12/25
2   Thanksgiving of 11/2008 was the best.   11/2008

With the added test case:

df['text'].str.extract('.*?(\d+/\d+/?\d*).*?|\
(January|February|March|April|May|June|July|August|September|October|November|December \d+)', expand = False)\
.fillna('').sum(1)

And you get

0       10/23/89
1    December 25
2        11/2008