Change the string to be replaced according to the matched string by replacing with Python regular expression

environment

Validated with Python 3.5.2.

Usage scene

When replacing with a regular expression, if you want to change the character string to be replaced according to the matched character string

As an example, you can do the following.

• I want to replace the number in the character string with the result of some operation. "abcd1234efg"-> "abcd2468efg" (double the number)
• I want to replace the numbers in the character string with Chinese numerals "abcd1234efg"-> "abcd 1,234 efg"

(I think it is possible to perform more complicated processing other than the above example)

Method

Use re.sub (), which is used for replacement with a normal regular expression, and a function that implements the processing you want to apply to the matched character string.

Specifically, it is re.sub (pattern, function, character string to be searched). If the argument of the function called above is matchobj, you can use the character string that matches the pattern by setting matchobj.group (0) in the function.

def replace(matchobj):
    s = matchobj.group(0)
    #What you want to do

re.sub(`hoge`, replace, "sampletext")

Implementation example

In the example below, we prepared a function toKansuji () that converts numbers to Chinese numerals and a function fixTax () that changes 5% consumption tax to 8%, and matched them in re.sub (). It is replaced with the result of applying the above function to the numbers.

code

sample.py

import re

def toKansuji(matchobj):
    a = ["one", "Ten thousand", "Billion", "Trillion", "Kyo"]
    b = ["one", "Ten", "hundred", "thousand"]
    c = ("", "one", "two", "three", "four", "Five", "Six", "Seven", "Eight", "Nine")
    s = matchobj.group(0)
    l = [int(i) for i in s]
    result = ""
    length = len(l)
    for i in range(length):
        n = l[length-1-i]
        if i%4 == 0 and i//4 > 0: 
            az = 1
            for j in range(4):
                if l[length-1-i-j] != 0: az = 0
            if az == 0: result = a[i//4] + result
        if n != 0:
            if i%4 != 0:
                result = b[i%4] + result
                if n != 1: result = c[n] + result
            else:
                result = c[n] + result
    return result

def fixTax(matchobj):
    d = int(matchobj.group(0))
    d = int(d / 1.05 * 1.08)
    return str(d)

if __name__ == '__main__':
    s1 = "315 yen including tax"
    print("Consumption tax 5%:", s1)
    print("Consumption tax 8%:", re.sub('\d+', fixTax, s1))

    s2 = "IPv4 is 4294967296"
    print("Numbers:", s2)
    print("Chinese numeral:", re.sub('\d{2,}', toKansuji, s2))

Execution result

Consumption tax 5%:315 yen including tax
Consumption tax 8%:324 yen including tax
Numbers:IPv4 is 4294967296
Chinese numeral:IPv4 is 492,494,967,296

The top two lines replace the consumption tax from 5% to 8%, and the bottom two lines replace the numbers with Chinese numerals.

Other

If you group using () in the pattern, you can also use the matched character string in that group by setting matchobj.group (n) in the function.

(Also, since both of the functions prepared this time are prepared for operation check, they may not return accurate values for all inputs.)

reference

https://docs.python.jp/3/library/re.html