How can I separate symbols [">", "<", ">=", "<="], numeric value and unit from a string by using regular expression in Python?

I would like to separate symbols, numeric value and unit from a list of string by using regular expression.

    x = ">=40.55%"

Currently I try to use the following regex in Python to separate symbol, numeric value, and unit

    match = re.findall(r'^(\A[>|<])*(\d+[.]+\d+)*([%|mg/dl|cm2]\Z)',i)

But, it doesn't give the expected output.

Expected output:

    symbol = >=
    value = 40.55
    unit = %

Can I know how to use regular expression in Python to separate a string into symbol, numeric value and unit?

Below I made some assumptions about your format, for example that numbers like .3 (to stand for 0.3) are disallowed.

import re

regex = re.compile(r'\A(=|<=|>=|<|>)(-?\d+(?:\.\d+)?)(%|mg/dl|cm2)\Z')

x = ">=40.55%"
m = regex.match(x)
symbol, value, unit = m.groups()
# symbol: '>='
# value: '40.55'
# unit: '%'

# let's try to match a different string: "=-345mg/dl"
regex.match("=-345mg/dl").groups()
# output: ('=', '-345', 'mg/dl')

Here, (?:...) denotes a non-capturing group. Note that regex-initial ^ and regex-final $ are respective synonyms of \A and \Z except in MULTILINE mode.

Check out the official Regular Expression HOWTO.

Credit goes to user Pranav Hosangadi for suggesting to match an optional minus sign to capture negative numbers.

Here's an answer that doesn't require you to pre-define the allowable units. Anything after the number is considered a unit.

Regex and explanation (Try online):

^([<>]=?)(-?\d+(?:\.\d+)?)(.*)$
-------------------------------
^                             $ : Start and end of string or line
 (      )(               )(  )  : Capturing groups for each portion of the string
  [<>]                          : Less than or greater than symbol
      =?                        : Optional equal symbol
          -?                    : Optional minus sign
            \d+                 : One or more digits
               (?:     )?       : Optional non-capturing group
                  \.\d+         : Decimal point followed by one or more digits
                           .*   : Any number of any character

Code:

result = re.findall(r"^([<>]=?)(-?\d+(?:\.\d+)?)(.*)$", ">=40.55%")
if result:
    symbol, value, unit = result[0]

which gives:

symbol = '>='
value = '40.55'
unit = '%'

A few notes about the pattern ^(\A[>|<])*(\d+[.]+\d+)*([%|mg/dl|cm2]\Z) why it does not give the expected output:

• This (\A[>|<])* optionally repeats a single char being one of > | < at the start of the string, which will only match a single char at the most. The \A should be at the start of the pattern, and you should repeat the character class inside of the capture group
• This [>|<] does not match a = char
• This [%|mg/dl|cm2] does match the single % but it does not mean matching alternatives with the |, it is a character class matching one of % | m etc...
• This (\d+[.]+\d+)* Matches a mandatory decimal part, but note that repeating a capture group captures the value of the last iteration. So if you would have for example this string >=40.55.2% the capture group value would be 5.2

Example using named capture groups:

import re

pattern = re.compile(r"\A(?P<symbol>[<>]=?)(?P<value>\d+(?:\.\d+)*)(?P<unit>%|mg/dl|cm2)\Z")
s = ">=40.55%"
m = pattern.match(s)
if m:
    print(m.groupdict())

Output

{'symbol': '>=', 'value': '40.55', 'unit': '%'}

The pattern explained:

\A(?P<symbol>[<>]=?)(?P<value>\d+(?:\.\d+)*)(?P<unit>%|mg/dl|cm2)\Z

• \A Start of string
• (?P<symbol>[<>]=?) Named group symbol, match one of < > and optional =
• (?P<value>\d+(?:\.\d+)*) Named group value, match 1+ digits and optionally repeat . and 1+ digits
• (?P<unit>%|mg/dl|cm2) Named group unit, match 1 of the alternatives
• \Z End of string

Regex demo | Python demo

Try

match = re.findall(r'(?:\A|\s)(=|<=|>=)(\d+\.\d+)(\%|mg/dl|cm2)(?:\Z|\s)',i)

matches xx.x

match = re.findall(r'(?:\A|\s)(=|<=|>=)(\d+(?:\.\d+)?)(\%|mg/dl|cm2)(?:\Z|\s)',i)

matches xx.x and xx, f.e. i = "something >=40.55% or =30cm2 etc." ==> result: [('>=', '40.55', '%'), ('=', '30', 'cm2')]