pyparsing: Trailing whitespace is not matched but still included for endloc calculation

user_3961349 0
Viewed 52

I am using Python 3.9.5 and pyparsing==3.0.6

I have a scenario where I need to match a word(alphanum) and optionally another word. However, pp.Optional() combined with trailing whitespace leads to incorrect endloc index.

import pyparsing as pp


def first_match(expression, text):
    for m, s, e in expression.scanString(text, maxMatches=1):
        return {'match': m, 'start': s, 'end': e}
    return None


expr = pp.Group(
    pp.Word(pp.alphanums) +
    pp.Optional(
        pp.Word(pp.alphanums)
    )
)

In the test bellow everything is as I expect. The entire string is matched from 0 to 3.

print(first_match(expr, "one"))
# {'match': ParseResults([ParseResults(['one'], {})], {}), 'start': 0, 'end': 3}

If I add trailing spaces, even though they are (correctly) not matched, they are included in the endloc index calculation. So the matched range is 0 to 8:

print(first_match(expr, "one     "))
# {'match': ParseResults([ParseResults(['one'], {})], {}), 'start': 0, 'end': 8}

Shouldn't the endloc index returned by scanString always be the index of the last match character?

pyparsing python
1 Answers

This is intended behavior, and is not a regression from pyparsing 2.4.7.

Think of Optional(expr) as "expr or empty". pyparsing will advance over intervening whitespace, and then, not finding the expression, consider the parse a successful match (since the expr it was looking for was not required).

To get the behavior you are looking for, you could use this format:

expr = pp.Group(
    pp.Word(pp.alphanums) + pp.Word(pp.alphanums)
    | pp.Word(pp.alphanums)
)

If there is a second alpha word present, it will parse them both, and if there is not, then it will just parse the first one.

Test with and without a second value:

print(first_match(expr, "one"))
print(first_match(expr, "one     "))
print(first_match(expr, "one     two"))
print(first_match(expr, "one     two    "))

Gives:

{'match': ParseResults([ParseResults(['one'], {})], {}), 'start': 0, 'end': 3}
{'match': ParseResults([ParseResults(['one'], {})], {}), 'start': 0, 'end': 3}
{'match': ParseResults([ParseResults(['one', 'two'], {})], {}), 'start': 0, 'end': 11}
{'match': ParseResults([ParseResults(['one', 'two'], {})], {}), 'start': 0, 'end': 11}
user_1652160
Related