regexp

ovince · 06-16-2007, 10:25 AM

Hi,

I would like to distinguish between files with name

1. *specIDs.dat

2. *_IDs.dat

So I wrote:

Code:

for fl in `find -type f -name "*tau[1-6][!0-9]*_IDs.dat" -print`; do
 echo $fl;
done

Why this does not work? When I put
"*tau[1-6][!0-9]*IDs.dat" it works but finds 2. also.

thanks
o.

ta0kira · 06-16-2007, 10:54 AM

Try using "-regex" instead of "-name". Also, "*" itself doesn't do anything; it means "previous character repeated 0 or more times". That means you should have a "." (any character) before the first "*":

Code:

.*tau[1-6][!0-9]*_IDs.dat

So is this what you intend?

Code:

(anything consisting of 0 or more chars)("tau")(single char 1 through 6)(0 or more of "!" or 0 through 9)("_IDs.dat")

ta0kira

edit: If you mean "[!0-9]*" as "any group of something besides numbers" you need to change "!" to "^".

wjevans_7d1@yahoo.co · 06-16-2007, 11:02 AM

Quote:

"*" itself doesn't do anything; it means "previous character repeated 0 or more times"

That's true for a regex, but not for -name.

ovince · 06-16-2007, 11:12 AM

I just have files with names:

file_tau1_IDs.dat
file_tau3_IDs.dat
file_tau5_IDs.dat
file_tau6_IDs.dat
so on ...

and I want to distinguish these from similar files

file_tau1_specIDs.dat
file_tau3_specIDs.dat
file_tau5_specIDs.dat
file_tau6_specIDs.dat
....

Numbers goes from 1 to 10

ta0kira · 06-16-2007, 11:27 AM

Code:

find -type f -regex "file_tau([1-9]|10)_IDs.dat"

Will that work? I don't seem to have luck with 'find' regex, so maybe you need to just 'egrep':

Code:

find -type f | egrep "/file_tau([1-9]|10)_IDs.dat$"

Quote:

Originally Posted by wjevans_7d1@yahoo.co

That's true for a regex, but not for -name.

Yes, I understand that. I was more pointing out that a regex isn't the same as a wildcard.
ta0kira

edit: forgot "tau"

ovince · 06-16-2007, 12:00 PM

i was so unprecise with definition what I would like to do. sorry.

I have this kind of files in the different sub directories:

file_tau1_IDs.dat
file_tau3_specIDs.dat
file_tau5_IDs.dat
file_tau6_specIDs.dat
file_tau10_IDs.dat
file_tau2_specIDs.dat
so on ...

Numers running from 1 to 10. I would like to manipulate only files that contain
numbers 1 to 6 (ie no 7, 8, 9, 10) and only that with _IDs.dat (ie. no *_specIDs.dat).
Especially I have a problem to exclude file that has 10.

slakmagik · 06-16-2007, 12:18 PM

find -regex '.*file_tau[1-6]_IDs.dat' | xargs someCmd
or just shell globbing would do for what you've just described:
someCmd file_tau[1-6]_IDs.dat

-- Oops. Need to read more carefully - shell globbing won't suffice if some are in subdirs. Though something like 'someCmd {,*/}file_tau[1-6]_IDs.dat' would work for one level of subdirs and so on.

Tinkster · 06-17-2007, 03:22 AM

In that case the -exec option to find may be helpful?

Cheers,
Tink

wjevans_7d1@yahoo.co · 06-17-2007, 10:54 AM

ovince, with all the suggestions and comments, I'd like to know: Is your original problem solved?