Reputation: 1255
I am using regular expressions in python to finds dates like 09/2010 or 8/1976 but not 11/12/2010. I am using the following lines of codes but it does not work in some cases.
r'([^/](0?[1-9]|1[012])/(\d{4}))'
Upvotes: 1
Views: 76
Reputation: 1255
After working on this problem I came to this solution:
This works very well!
df['text'].str.extractall(r'(?P<Date>(?P<month>\d{1,2})/?(?P<day>\d{1,2})?/(?P<year>\d{2,4}))')
Upvotes: 0
Reputation: 195438
This, a little bit explicit code, uses re.sub
and datetime.strptime
to parse/validate the input string:
import re
import datetime
s = '09/2010, 8/1976, 11/8/2010, 09/06/15, 12/1987, 13/2011, 09/13/2001'
r = re.compile(r'\b(\d{1,2})/(?:(\d{1,2})/)?(\d{2,4})\b')
def validate_date(g, parsed_values):
if not g.group(2) is None:
s = '{:02d}/{:02d}/{:04d}'.format(*map(int, g.groups()))
else:
s = '01/{:02d}/{:04d}'.format(int(g.group(1)), int(g.group(3)))
try:
datetime.datetime.strptime(s, '%d/%m/%Y')
parsed_values.append(g.group())
return
except:
pass
parsed_values = []
r.sub(lambda g: validate_date(g, parsed_values), s)
print(parsed_values)
Prints:
['09/2010', '8/1976', '11/8/2010', '09/06/15', '12/1987']
EDIT: Shortened the code.
Upvotes: 1
Reputation: 1598
import re
rgx = "(?:\d{1,2}\/)?\d{1,2}\/\d{2}(?:\d{2})?"
dates = "09/2010, 8/1976, 11/12/2010, 09/06/15 .."
result = re.findall(rgx, dates)
print(result)
# ['09/2010', '8/1976', '11/12/2010', '09/06/15']
Upvotes: 1