Multiple regex replacements with pandas

Question

I have a pandas series of various age ranges:

s = pd.Series([14,1524,2534,3544,65])

I would like to create a new series like this:

I can do this using mapping:

s = pd.Series([14,1524,2534,3544,65])
age_map = {
    14: '0-14',
    1524: '15-24',
    2534: '25-34',
    3544: '35-44',
    4554: '45-54',
    5564: '55-64',
    65: '65+'
}
s.map(age_map)

Also, using multiple regexes:

s = pd.Series([14,1524,2534,3544,65])
s = s.astype(str).str.replace(r'(\d\d)(\d\d)', r'\1-\2',regex=True)
s = s.astype(str).str.replace(r'14', r'0-14',regex=True)
s = s.astype(str).str.replace(r'65', r'65+',regex=True)
s

Question
Can we combine all three regexes into one advanced regex and obtain the same result?

something like:

s = pd.Series([14,1524,2534,3544,65])
pat = ''
pat_sub = ''
s = s.astype(str).str.replace(pat, pat_sub,regex=True)
s

cs95 · Accepted Answer · 2019-01-05 18:16:46Z

3

You can use a single callback function to handle all the cases:

def parse_str(match):
    a, b = match.groups()
    if not b:
        return f'0-{a}' if a == '14' else f'{a}+'    
    return f'{a}-{b}'

s.astype(str).str.replace(r'(\d{2})(\d{2})?', parse_str)

0     0-14
1    15-24
2    25-34
3    35-44
4      65+
dtype: object

This should work assuming your Series contains only either two or four digit numbers.

answered Jan 5, 2019 at 18:16

cs95

406k106 gold badges744 silver badges797 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Karn Kumar Over a year ago

Nice one! +1 for this syntax

cs95 Over a year ago

@pygo Thanks :-)

Karn Kumar Over a year ago

Ah, i see nice one though .. i'm adding this function in my note to future reference though.

BhishanPoudel · Accepted Answer · 2019-01-05 18:20:32Z

1

I liked the answer of @coldspeed which is more flexible and function is reusable.

However, I came up with pandas chain operation like this:

s = s.astype(str).str.replace(r'14', r'0-14',regex=True)
                 .str.replace(r'65', r'65+',regex=True)
                 .str.replace(r'(\d\d)(\d\d)', r'\1-\2',regex=True))

s

answered Jan 5, 2019 at 18:20

BhishanPoudel

17.4k27 gold badges123 silver badges190 bronze badges

Collectives™ on Stack Overflow

Multiple regex replacements with pandas

2 Answers 2

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related