Demonstration of regular expression (?(DEFINE) ...) for declaring resuable sub-patterns

regex_define.py

#!/usr/bin/env python3

import regex

PATTERN = regex.compile(r'''
    # Declare reusable sub-patterns
    (?(DEFINE)
        (?<COUNT>[0-9]+)
        (?<CURRENCY>EUR|GBP|USD)
        (?<QUANTITY>[0-9]+[.][0-9]+)
        (?<SEP>[\t ])
    )

    # The pattern to match
    (?<name>[A-Za-z]+)
    (?&SEP)+
    (?<currency>(?&CURRENCY))
    (?&SEP)
    (?<price>(?&QUANTITY))
    (?&SEP)
    (?<discount_pct>(?&QUANTITY))%
    (?&SEP)
    (?<quantity>(?&COUNT))
    ''',
    regex.VERBOSE,
)

m = PATTERN.match('Pencils\t\tEUR 0.10  5.00%  1000')
print({k: v for k, v in m.groupdict().items() if k == k.lower()})

$ python3 regex_define.py
{'name': 'Pencils', 'currency': 'EUR', 'price': '0.10', 'discount_pct': '5.00', 'quantity': '100'}

• sub-pattern names don't have to be uppercase, but I find it a handy convention
• sub-pattern names are included in m.groupdict(), hence the extra filtering
• (?(DEFINE) is not supported by Python's stdlib re module. This uses https://pypi.org/project/regex/