regular expression : the dollar sign ($) work with re.match() or re.search() ？

[iMath]

I only know the dollar sign ($) will match a pattern from the
end of a string,but which method does it work with ,re.match() or re.search() ？

 

[Jussi Piitulainen]

I only know the dollar sign ($) will match a pattern from the end of
a string, but which method does it work with, re.match() or
re.search()

It works with both. With re.match, the pattern has to match at the
start of the string _and_ the $ has to match the end of the string (or
a line); re.search scans the string until it finds a suitable start.

What was the weird character that you used as a question mark? I
removed them because they confuse the newsreader I use.

 

[Chris Angelico]

What was the weird character that you used as a question mark? I
removed them because they confuse the newsreader I use.

It appears to be Unicode Character 'FULLWIDTH QUESTION MARK' (U+FF1F).
Normally I'd be inclined to simply use U+003F instead, but hey, it's a
question mark still.

ChrisA

 

[Peter Otten]

I only know the dollar sign ($) will match a pattern from the
end of a string,but which method does it work with ,re.match() or
re.search() ？

Why not try it out in the interactive interpreter? Here's the "deluxe
version":

def demo(pattern="mid$", texts=["start mid end", "start mid", "mid end",

"mid"], matchers=[re.match, re.search]):
.... print "pattern:", pattern
.... for text in texts:
.... for matcher in matchers:
.... name = matcher.__name__
.... print u"\N{CHECK MARK}" + name if matcher(pattern,
text) else (" "*(len(name)+1)),
.... print repr(text)
....pattern: mid$
'start mid end'
✓search 'start mid'
'mid end'
✓match ✓search 'mid'

 

[Jussi Piitulainen]

It shows fine in my news reader, perhaps you should consider changing to
one that works properly (btw I am using pan on a fedora 17 netbook)

I was just curious why anyone would use anything other than the ASCII
question mark as an ordinary question mark when writing in English in
a newsgroup.

The post had this:

0000520 61 72 63 68 28 29 20 20 ef bc 9f 0a
a r c h ( ) sp sp o < us nl

od is showing (ef bc 9f) as (o < us) but since they are not individual
characters anyway, never mind that. Google tells me (ef bc 9f) is
UTF-8 for U+FF1F FULLWIDTH QUESTION MARK, so now I basically have my
answer as to what it is, though still not as to why one would use it.

The ordinary question mark would look like this:

0000000 61 72 63 68 28 29 20 3f 0a
a r c h ( ) sp ? nl

 

[Ian Kelly]

I guess you can consider re.match's pattern to be
prefixed with '^'.

You can in this case, but they're not equivalent in multi-line mode:
<_sre.SRE_Match object at 0x0475BFA8>

 

[iMath]

在 2012å¹´9月26日星期三UTC+8下åˆ3æ—¶38分50秒，iMath写é“：

I only know the dollar sign ($) will match a pattern from the

end of a string,but which method does it work with ,re.match() or re.search() ？

I thought re.match('h.$', 'hbxihi') will match ‘hi’ ,but itdoes not .so why ?

 

[Steven D'Aprano]

在 2012å¹´9月26日星期三UTC+8下åˆ3æ—¶38分50秒，iMath写é“：

I thought re.match('h.$', 'hbxihi') will match ‘hi’ ,but it does not .so
why ?

re.match only matches at the *start* of the string, so "h.$" tries to
match:

* start of string
* literal h
* any character
* end of string


You want re.search, which will search the entire string and match "hi" at
the end of the string.