一个数字加上两个字符的正则表达式，如“1st”、“2nd”、“10th”、“22nd”？

Question

I have a dataset of phone calls transcribed into text, where each sample contains text. I'm trying to identify all the samples where dates are mentioned. To be clear, I'm only looking for samples where a number and two additional characters are present, like "1st", "2nd", "25th".

Right now, I have a rather brute force approach of going about it. It does the job, but is there a cleaner way to achieve the same using regex?

def date_mentioned(text):
    date_list = ['1st', '2nd', '3rd', '4th', '5th', '6th', '7th', '8th', '9th', '10th', '11th', '12th', '13th', '14th', '15th', '16th', '17th', '18th', '19th', '20th', '21st', '22nd', '23rd', '24th', '25th', '26th', '27th', '28th', '29th', '30th', '31st']

    for date in date_list:
        if re.search(date, text):
            return True
    return False

Answer 1

You could use a regular expression for this. You may try with:

r'\d{1,2}(?:st|nd|rd|th)'

See demo

---

• \\d{1,2}(?:st|nd|rd|th)
  • \\d{1,2} matches a digit (equal to [0-9] )
    • {1,2} Quantifier — Matches between 1 and 2 times
  • Non-capturing group (?:st|nd|rd|th)
    • 1st Alternative st
      • st matches the characters st literally (case sensitive)
    • 2nd Alternative nd
      • nd matches the characters nd literally (case sensitive)
    • 3rd Alternative rd
      • rd matches the characters rd literally (case sensitive)
    • 4th Alternative th
      • th matches the characters th literally (case sensitive)tive)

Answer 2

For general numbers, \\d*([02-9]1st|2nd|3rd|([04-9]|1[1-3])th) should do what you want. You can restrict the numbers further for dates, but full validation is complex (months, leapyears, etc), so I'd recommend just blindly parsing the number and then validating it afterwards.

Edit: Thanks for pointing out the mistake with 3rd; fixed.

Answer 3

You find these dates with:

[0-9]{1,2}(?:st|nd|rd|th)

Explanation:
1 or 2 digits,
followed by st, nd, rd or th

Answer 4

Since you're looking for ordinal numbers, rules are:

If the number ends with 1 and is not 11, add 'st'
If the number ends with 2 and is not 12, add 'nd'
If the number ends with 3 and is not 13, add 'rd'
for all the other numbers, add 'th'

A regex that can distinguish between this cases is:

'^11th|12th|13th|\d*(1st|2nd|3rd|[04-9]th)$'

And the application is:

def date_mentioned(text):
    if re.match('^11th|12th|13th|\d?(1st|2nd|3rd|[04-9]th)$', text):
        return True
    return False

---

RegEx explanation
We're looking for this sequence:

^ : Start of the string
11th : string 11th
| : or
12th : string 12th
| : or
13th : string 13th
| : or
\\d? : 0 or 1 digits ( : followed by

1st : string 1st
| : or
2nd : string 2nd | : or
3rd : string 3rd | : or
[04-9] : one number between 0 and the comprehensive range 4-9
th : string th
) : closing 'followed by'
$ : end of the string