Regular Expression for date in python?

Question

How to check date value in pandas dataframe through regex and pick wrong date type values?

date: df

Date_col
22-01-2016
2016-01-22
2016/01/22
2018-12-25 09:27:53
22-Jan-2016
abcd
1203
2072006
20030201

output:

Bad_Date
22-Jan-2016
abcd
1203
2072006
20030201

may be regular expression which allows only numbers,'-' ,'/', ':' would work out this problem.

Are you looking to verify whether a column simply has a specific date-like format or are you looking to validate correct dates as well? — MonkeyZeus, Dec 23 '19 at 14:44
For example, is `2019-02-30` a bad date? What about `2018-45-12`? — MonkeyZeus, Dec 23 '19 at 14:45
It should contain proper date format either 22-07-2019 or 2019-07-22 other than this it would be a bad date. — Mohekar, Dec 23 '19 at 16:17
I am looking for regular expression which only allows combination of numbers[0-9] with (-,/,:). I think solve my problem. Other than this is a bad date. — Mohekar, Dec 23 '19 at 16:20

score 0 · Accepted Answer · answered Dec 23 '19 at 14:37

0

One idea could be to see which values pandas.to_datetime rejects:

import pandas as pd

bad_times = []
for i in df.Date_col.values:
    try:
        pd.to_datetime(i)
    except ValueError:
        bad_times.append(i)

answered Dec 23 '19 at 14:37

ignoring_gravity

6,677
4
32
65

score 0 · Answer 2 · answered Dec 23 '19 at 15:25

Use str.contains with regular expressions:

wrong_dates1 = ~df['Date_col'].str.contains('[-/:]')   # rows without the characters -, /, :
wrong_dates2 = df['Date_col'].str.contains('[A-Za-z]') # rows with letters

df[wrong_dates1 | wrong_dates2]

      Date_col
4  22-Jan-2016
5         abcd
6         1203
7      2072006
8     20030201

side note

If you would consider 22-Jan-2016 and 20030201 as valid dates (which they are), you can simply use pd.to_datetime with errors='coerce' which casts invalid dates tot NaT:

bad_dates = pd.to_datetime(df['Date_col'], errors='coerce').isna()
df[bad_dates]

  Date_col
5     abcd
6     1203
7  2072006

Regular Expression for date in python?

2 Answers2