Stripping whitespace in a find command

kj6loh · 12-11-2009, 05:13 PM

I don't want to strip the beginning or the end.
but I want to do something like

find <dir> -iname *<1wd><spaces><2wd>*

I've tried
find <dir> -iname *1wd\ +2wd*
find <dir> -iname *1wd[\ ]+2wd*
find <dir> -iname *1wd[\ ]\+2wd*
to no avail.

Komakino · 12-11-2009, 05:27 PM

You want something along the lines of:

Code:

find <dir> -iregex "*wd\ *2wd"

but the -iregex syntax is odd because it matches on the whole file name, even if you've already specified a directory.
e.g. the file:

/folder/afile.bat

will match with

-iregex "folder/a.*\.bat"

but not

-regex "a.*\.bat"

(and any number of spaces would be "\ *")

ghostdog74 · 12-11-2009, 07:04 PM

do it like this

Code:

find /path -type f -iname "* *"

kj6loh · 12-12-2009, 08:23 AM

ghostdog yours would not work and I tell you why that gets every file with a space.

Let me give you guys a more concrete example.

I have 11 files

Code:

aaaaab  caaaa
aaaad daaaaaa
aaab   caaaa
aae   faa
aaaaaaab               c
aab       ccccaaa
ddddd
b    c
aaaaC   B
B      Caaaa
baaac

I want to get all the files with b then c regardless of the number of spaces in between and regardless of case. So this is the output:

Code:

aaaaab  caaaa
aaab   caaaa
aaaaaaab               c
aab       ccccaaa
b    c
B      Caaaa

ghostdog74 · 12-12-2009, 08:28 AM

then add your "b"s and "c"s into the expression. Its easy, so I leave it to you to experiment. the command line is there for you to use.

colucix · 12-12-2009, 08:45 AM

Why not simply a good-ol' pipe?

Code:

find . -type f | grep -E '.*[bB][ ]+[cC].*'

catkin · 12-12-2009, 08:54 AM

Quote:

Originally Posted by kj6loh

I want to get all the files with b then c regardless of the number of spaces in between and regardless of case.

Then you want

Code:

find /path -type f -iname "wrd1* *wrd2"

or maybe

Code:

find /path -type f -iname "*wrd1* *wrd2*"

ghostdog74 · 12-12-2009, 09:04 AM

Quote:

Originally Posted by colucix

Why not simply a good-ol' pipe?

Code:

find . -type f | grep -E '.*[bB][ ]+[cC].*'

because you are not going to call grep for every file it finds, even the ones that are not needed?? this produces overhead. Its better to let find do the walk and talk.

Code:

find . -type f <for OP to find out>  ".*b[ ]+c.*"

colucix · 12-12-2009, 09:09 AM

Putting together suggestions by Komakino, ghostdog74 and catkin, most likely this is what you're looking for:

Code:

find . -type f -iname "*b* *c*" -iregex '.*b[ ]+c.*'

the output will not contain matching directories (see note by Komakino).

colucix · 12-12-2009, 09:10 AM

Quote:

Originally Posted by ghostdog74

because you are not going to call grep for every file it finds, even the ones that are not needed?? this produces overhead.

That's true. I stand corrected.

ghostdog74 · 12-12-2009, 09:10 AM

Quote:

Originally Posted by catkin

Then you want

Code:

find /path -type f -iname "wrd1* *wrd2"

or maybe

Code:

find /path -type f -iname "*wrd1* *wrd2*"

according to OP, that's not his requirement. he wants to find multiple spaces as well..

ghostdog74 · 12-12-2009, 09:12 AM

Quote:

Originally Posted by colucix

Putting together suggestions by Komakino, ghostdog74 and catkin, most likely this is what you're looking for:

Code:

find . -type f -iname "*b* *c*" -iregex '.*b[ ]+c.*'

the output will not contain matching directories (see note by Komakino).

don't need -iname

catkin · 12-12-2009, 09:27 AM

Quote:

Originally Posted by ghostdog74

according to OP, that's not his requirement. he wants to find multiple spaces as well..

The * pattern elements also match spaces.

Code:

c:/tmp$ touch 'a   b'
c:/tmp$ find . -iname 'a* *b'
./a   b

ghostdog74 · 12-12-2009, 09:35 AM

Quote:

Originally Posted by catkin

The * pattern elements also match spaces.

Code:

c:/tmp$ touch 'a   b'
c:/tmp$ find . -iname 'a* *b'
./a   b

i originally wanted to give him this

Code:

find . -iname "*b* *c*"

but i am considering the fact that he might have files like

Code:

blahb     blah    cblahblah

so, anyway, best to let OP decide

catkin · 12-12-2009, 10:00 AM

Quote:

Originally Posted by ghostdog74

i originally wanted to give him this

Code:

find . -iname "*b* *c*"

but i am considering the fact that he might have files like

Code:

blahb     blah    cblahblah

so, anyway, best to let OP decide

Sorry -- you are right; I misunderstood the requirement which is clearly expressed by example in this post