Unix Problems

RobbieM

I'm using SUSE csh

I have a unix problem.

I want a command line to search a unix hierarchy for files with a pipe “|”in their name and rename them to dash “-”.

any ideas?

RobbieM

This is what I'm testing so far...
find ../ | egrep '\|' | xargs -I {} -e echo cp '"{}"' \`echo '"{}"' \| sed "'s/|/-/g'" \`
It's close. eventually change the "echo cp" for "mv"

Sparks

Moved to the Unix forum.

Tillotson

There's probably a much nicer way to do this using the `-exec` option or xargs.

for old in $( find test/* -name "*|*"  -type f )
do
    new="$( echo $old | sed 's/|/-/' )"
    mv $old $new
done

Hopefully someone who knows what they're doing will come along soon.

Skrynesaver

tested, it works only assumption is that you shoot users who put spaces in file names, also assuming you will henceforth shoot users who put pipes in file names, but that's not necessary for the script to work

for i in $(find ./files -name \*\|\* -exec echo {} \;) ;do 
mv  $i $(echo $i| sed 's/|/-/g')
done

niallb

Skrynesaver wrote: »

... you shoot users who put spaces in file names, also assuming you will henceforth shoot users who put pipes in file names, but that's not necessary for the script to work ...

It's just necessary.

Capt'n Midnight

Skrynesaver wrote: »

also assuming you will henceforth shoot users who put pipes in file names,

and in the microsoft world they've stuck in &'s and +'s and reserved devices like .con

fairly unparsable on the command line be glad you're not working with windows


in the windows world however, you can fall back to 8.3 names to avoid spaces and some dodgy characters - is there something similar in linux/unix or should one just use something like sed and pick a delimiter you know isn't in use anywhere ?

Naikon

Best to use xargs where the param list is large. This is because -exec runs a seperate instance of the command to be run for each match.

RobbieM

%> set tag = '"*log.gz"' ; set fin = "find . -name $tag" ; echo "$fin" ; pwd ; $fin
find . -name "*log.gz"
/tmp/test1
find: No match.
%> find . -name "*log.gz"
./testfile.log.gz
./testfil2.log.gz
./testfil3.log.gz

Why is there no match within the command line, but the find works following the command line?

RobbieM

Stuck on this... This is the biggest challenge I have come up against. It will take a great programmer to figure this one out !

here it is:

take a file list (with some multi element lines) and reformat into stacked columns for a csv.

example:



""Similar problem. I revisited this again recently but to no avail.

all_errors
bin_error exe_error noa_error_error noa
bin_error exe_error test_error noa tap tag_error
bin_error exe_error noa tap tag_error lvis
bin_error
exe_error
test_error
noa
to


[Embedded Image Removed]

some errors in my paste above... basically list to stacked cols.

It’s impossible!!!!""


I will post my solution if I ever find one. Trying to do this in csh, and therein lies the problem. Going to learn Perl or Ruby or tcl to try solve.




| noa_error | test_error |
| exec_errors | exec_errors | exec_errors
all_errors | bin_error | bin_error | bin_error | bin_error exe_error | test_error | noa |

mobitron

RobbieM wrote: »

%> set tag = '"*log.gz"' ; set fin = "find . -name $tag" ; echo "$fin" ; pwd ; $fin
find . -name "*log.gz"
/tmp/test1
find: No match.
%> find . -name "*log.gz"
./testfile.log.gz
./testfil2.log.gz
./testfil3.log.gz

Why is there no match within the command line, but the find works following the command line?

Because csh is useless for this kind of thing. It works OK for interactive commands that you only want to run once. If you try doing anything (even basic stuff) with variables, redirection or filename matching it gets hopelessly confused.

If you're trying to create any kind of useful shell script, use ksh or bash.

To do what you want in ksh:

tag="*log.gz"; echo $tag; fin="find . -name $tag";echo $fin; pwd; $fin
*log.gz
find . -name *log.gz
/appl
./common/testfile2.log.gz
./common/testfile3.log.gz
./common/testfile.log.gz