Python正则表达式匹配组具有超过预期的对象

Question

I am using re.search to parse a line for 3 separate pieces of data. (date temperature and pressure) The line looks like this.

line= "2015-10-08-22-50   27.3   1015.03"

I want to use pattern matching so that I can be very robust against malformed lines. Using split has failed for that reason.

I built following re.

m= re.search("^(2\d{3}-\d{2}-\d{2}-\d{2}-\d{2})\s+(\d+.\d+)\s+(\d+.\d+)$", line)

The parsing is fine however the match groups surprised me.

>>> m.groups(1)
('2015-10-08-23-00', '27.3', '1014.99')
>>> m.groups(2)
('2015-10-08-23-00', '27.3', '1014.99')
>>> m.groups(3)
('2015-10-08-23-00', '27.3', '1014.99')

I (naively) had expected.

>>> m.groups(1)
('2015-10-08-23-00')
>>> m.groups(2)
('27.3')
>>> m.groups(3)
('1014.99')

For now I work around this by using indices.

dt= m.groups(1)[0]
t = m.groups(2)[1]
p = m.groups(3)[2]

I conclude that the re that I believed was OK must be flawed or not as clean as possible.

What's missing?

Thanks, Gert

Answer 1

Instead of:

m.groups(1)

I think you want:

m.groups()[0]

The parameter to groups() is a default value, not a position in the tuple it returns. So you don't need to pass it anything. You do need to index the tuple it returns.

help(m.groups)
Help on built-in function groups:

groups(...)
    groups([default=None]) -> tuple.

    Return a tuple containing all the subgroups of the match, from 1.
    The default argument is used for groups
    that did not participate in the match

Answer 2

to capture parenthesized subgroup
use group, not groups

print(m.group(1))

2015-10-08-22-50

print(m.group(2))
27.3

print(m.group(3))
1015.03

print(m.group(1,3))

('2015-10-08-22-50', '1015.03')

Answer 3

The argument to m.groups() is not which capture group to return. It's an optional default value to use for any capture groups that didn't match anything. Either way, the function returns a list containing all the capture groups, and you have to index it to get a particular one.

Answer 4

There is nothing wrong with your pattern.

You could use named groups to clarify:

>>> pat=re.compile(r"""^(?P<date>2\d{3}-\d{2}-\d{2}-\d{2}-\d{2})\s+
...                     (?P<temp>\d+.\d+)\s+
...                     (?P<pres>\d+.\d+)$""", re.X)
>>> line= "2015-10-08-22-50   27.3   1015.03"
>>> m=pat.search(line)

Which then produces a dictionary:

>>> m.groupdict()
{'date': '2015-10-08-22-50', 'temp': '27.3', 'pres': '1015.03'}
>>> m.group('date')
'2015-10-08-22-50'
>>> m.group('temp')
'27.3'

But can also be accessed as usual:

>>> m.group(1)
'2015-10-08-22-50'
>>> m.group(2)
'27.3'

Or index the tuple returned by groups() but I don't think that is very clear (as you seem to have discovered...)

>>> m.groups()[2]
'1015.03'

If you use named groups, you get the best of all worlds.