regex problem with re and fnmatch

[Fabian Braennstroem]

Hi,

I would like to use re to search for lines in a files with
the word "README_x.org", where x is any number.
E.g. the structure would look like this:
[[file:~/pfm_v99/README_1.org]]

I tried to use these kind of matchings:
# org_files='.*README\_1.org]]'
org_files='.*README\_*.org]]'
if re.match(org_files,line):

Unfortunately, it matches all entries with "README.org", but
not the wanted number!?

After some splitting and replacing I am able to check, if
the above file exists. If it does not, I start to search for
it using the 'walk' procedure:

for root, dirs, files in
os.walk("/home/fab/org"):
for name in dirs:
dirs=os.path.join(root, name) + '/'
for name in files:
files=os.path.join(root, name)
if fnmatch.fnmatch(str(files), "README*"):
print "File Found"
print str(files)
break

As soon as it finds the file, it should stop the searching
process; but there is the same matching problem like above.
Does anyone have any suggestions about the regex problem?
Greetings!
Fabian

[John Machin]

On Nov 21, 8:05 am, Fabian Braennstroem <[e-mail]:

Hi,

I would like to use re to search for lines in a files with
the word "README_x.org", where x is any number.
E.g. the structure would look like this:
[[file:~/pfm_v99/README_1.org]]

I tried to use these kind of matchings:
# org_files='.*README\_1.org]]'
org_files='.*README\_*.org]]'
if re.match(org_files,line):

First tip is to drop the leading '.*' and use search() instead of
match(). The second tip is to use raw strings always for your
patterns.

>
Unfortunately, it matches all entries with "README.org", but
not the wanted number!?

\_* matches 0 or more occurrences of _ (the \ is redundant). You need
to specify one or more digits -- use \d+ or [0-9]+

The . in .org matches ANY character except a newline. You need to
escape it with a \.

>>pat = r'README_\d+\.org'
re.search(pat, 'xxxxREADME.org')
re.search(pat, 'xxxxREADME_.org')
re.search(pat, 'xxxxREADME_1.org')

<_sre.SRE_Match object at 0x00B899C0>

>>re.search(pat, 'xxxxREADME_9999.org')

<_sre.SRE_Match object at 0x00B899F8>

>>re.search(pat, 'xxxxREADME_9999Zorg')

>
After some splitting and replacing I am able to check, if
the above file exists. If it does not, I start to search for
it using the 'walk' procedure:

I presume that you mean something like: """.. check if the above file
exists in some directory. If it does not, I start to search for it
somewhere else ..."""

>
for root, dirs, files in
os.walk("/home/fab/org"):

for name in dirs:
dirs=os.path.join(root, name) + '/'

The above looks rather suspicious ...
for thing in container:
container = something_else
????
What are you trying to do?

for name in files:
files=os.path.join(root, name)

and again ....

if fnmatch.fnmatch(str(files), "README*"):

Why str(name) ?

print "File Found"
print str(files)
break

fnmatch is not as capable as re; in particular it can't express "one
or more digits". To search a directory tree for the first file whose
name matches a pattern, you need something like this:
def find_one(top, pat):
for root, dirs, files in os.walk(top):
for fname in files:
if re.match(pat + '$', fname):
return os.path.join(root, fname)

As soon as it finds the file,

"the" file or "a" file???

Ummm ... aren't you trying to locate a file whose EXACT name you found
in the first exercise??

def find_it(top, required):
for root, dirs, files in os.walk(top):
if required in files:
return os.path.join(root, required)

it should stop the searching
process; but there is the same matching problem like above.

HTH,
John

[Fabian Braennstroem]

Hi John,

John Machin schrieb am 11/20/2007 09:40 PM:

On Nov 21, 8:05 am, Fabian Braennstroem <[e-mail]:

>Hi,

I would like to use re to search for lines in a files with
the word "README_x.org", where x is any number.
E.g. the structure would look like this:
[[file:~/pfm_v99/README_1.org]]

I tried to use these kind of matchings:
# org_files='.*README\_1.org]]'
org_files='.*README\_*.org]]'
if re.match(org_files,line):

First tip is to drop the leading '.*' and use search() instead of
match(). The second tip is to use raw strings always for your
patterns.

>Unfortunately, it matches all entries with "README.org", but
not the wanted number!?

\_* matches 0 or more occurrences of _ (the \ is redundant). You need
to specify one or more digits -- use \d+ or [0-9]+

The . in .org matches ANY character except a newline. You need to
escape it with a \.

>>>pat = r'README_\d+\.org'
re.search(pat, 'xxxxREADME.org')
re.search(pat, 'xxxxREADME_.org')
re.search(pat, 'xxxxREADME_1.org')

<_sre.SRE_Match object at 0x00B899C0>

>>>re.search(pat, 'xxxxREADME_9999.org')

<_sre.SRE_Match object at 0x00B899F8>

>>>re.search(pat, 'xxxxREADME_9999Zorg')

Thanks a lot, works really nice!

>After some splitting and replacing I am able to check, if
the above file exists. If it does not, I start to search for
it using the 'walk' procedure:

I presume that you mean something like: """.. check if the above file
exists in some directory. If it does not, I start to search for it
somewhere else ..."""

> for root, dirs, files in
os.walk("/home/fab/org"):

> for name in dirs:
dirs=os.path.join(root, name) + '/'

The above looks rather suspicious ...
for thing in container:
container = something_else
????
What are you trying to do?

> for name in files:
files=os.path.join(root, name)

and again ....

> if fnmatch.fnmatch(str(files), "README*"):

Why str(name) ?

> print "File Found"
print str(files)
break

fnmatch is not as capable as re; in particular it can't express "one
or more digits". To search a directory tree for the first file whose
name matches a pattern, you need something like this:
def find_one(top, pat):
for root, dirs, files in os.walk(top):
for fname in files:
if re.match(pat + '$', fname):
return os.path.join(root, fname)

>As soon as it finds the file,

"the" file or "a" file???

Ummm ... aren't you trying to locate a file whose EXACT name you found
in the first exercise??

def find_it(top, required):
for root, dirs, files in os.walk(top):
if required in files:
return os.path.join(root, required)

Great :-) Thanks a lot for your help... it can be so easy :-)
Fabian