Extracting float or int number and substring from a string

Question:

I’ve just learned regex in python3 and was trying to solve a problem.
The problem is something like this:

You have given a string where the first part is a float or integer number and the next part is a substring. You must split the number and the substring and return it as a list. The substring will only contain the alphabet from a-z and A-Z. The values of numbers can be negative.
For example:

  1. Input: 2.5ax
    Output:[‘2.5′,’ax’]
  2. Input: -5bcf
    Output:[‘-5′,’bcf’]
  3. Input:-69.67Gh
    Output:[‘-69.67′,’Gh’]

and so on.

I did several attempts with regex to solve the problem.

1st attempt:

import re
i=input()
print(re.findall(r'^(-?d+(.d+)?)|[a-zA-Z]+$',i))

For the input -2.55xy, the expected output was [‘-2.55′,’xy’]
But the output came:

[(‘-2.55’, ‘.55’), (”, ”)]

2nd attempt:
My second attempt was similar to my first attempt just a little different:

import re
i=input()
print(re.findall(r'^(-?(d+.d+)|d+)|[a-zA-Z]+$',i))

For the same input -2.55xy, the output came as:

[(‘-2.55’, ‘2.55’), (”, ”)]

3rd attempt:
My next attempt was like that:

import re
i=input()
print(re.findall(r'^-?[1-9.]+|[a-z|A-Z]+$',i))

which matched the expected output for -2.55xy and also with the sample examples. But when the input is 2..5 or something like that, it considers that also as a float.

4th attempt:

import re
i=input()
value=re.findall(r"[a-zA-Z]+",i)
print([i.replace(value[0],""),value[0]])

which also matches the expected output but has the same problem as 3rd one that goes with it. Also, it doesn’t look like an effective way to do it.

Conclusion:
So I don’t know why my 1st and 2nd attempt isn’t working. The output comes with a list of tuples which is maybe because of the groups but I don’t know the exact reason and don’t know how to solve them. Maybe I didn’t understand the way the pattern works. Also why the substring didn’t show in the output?
In the end, I want to know what’s the mistake in my code and how can I write better and more efficient code to solve the problem. Thank you and sorry for my bad English.

Asked By: Samsil Arefeen

||
Source

Answers:

The alternation | matches either the left part or the right part.

If the chars a-zA-Z are after the digit, you don’t need the alternation | and you can use 2 capture groups to get the matches in that order.

Then using re.findall will return a list of tuples for the capture group values.

(-?d+(?:.d+)?)([a-zA-Z]+)

Explanation

  • ( Capture group 1
    • -?d+ Match an optional –
    • (?:.d+)? Optionally match . and 1+ digits using a non capture group (so it is not outputted separately by re.findall)
  • ) Close group 1
  • ( Capture group 2
    • [a-zA-Z]+ Match 1+ times a char a-z or A-Z
  • ) Close group 2

regex demo

import re

strings = [
    "2.5ax",
    "-5bcf",
    "-69.67Gh",
]

pattern = r"(-?d+(?:.d+)?)([a-zA-Z]+)"
for s in strings:
    print(re.findall(pattern, s))

Output

[('2.5', 'ax')]
[('-5', 'bcf')]
[('-69.67', 'Gh')]
Answered By: The fourth bird

lookahead and lookbehind in re.sub simplify things sometimes.

  • (?<=d) look behind
  • (?=[a-zA-Z]) look ahead

that is split between the digit and the letter.

strings = [
    "2.5ax",
    "-5bcf",
    "-69.67Gh",
]

for s in strings:
    print(re.split(r'(?<=d)(?=[a-zA-Z])', s))


['2.5', 'ax']
['-5', 'bcf']
['-69.67', 'Gh']
Answered By: LetzerWille