Shell Programming and Scripting

I have a requirement where I need to move files that are being written too. I am unable to use sed since I believe this will stop the files from getting written. There is a directory containing about 50 unique files, the date is the same on all of them.

I need to change the date string on each file name to yesterday. So example:

Code:

224.0.26.10_10017_20120112.mcrec
224.0.26.10_10017_20120112.mcrec.idx
224.0.26.102_10044_20120112.mcrec
224.0.26.102_10044_20120112.mcrec.idx

Needs to become:

Code:

224.0.26.10_10017_20120111.mcrec
224.0.26.10_10017_20120111.mcrec.idx
224.0.26.102_10044_20120111.mcrec
224.0.26.102_10044_20120111.mcrec.idx

I know I can try a loop of some sort. I just don't know how to target just the date since the files are unique.

Code:

dateYest=$(date -d yesterday +%Y%m%d)

for i in /dir/* 
do

mv SomeMagic > MoreMagic$dateYest 

done.

Maybe I am wrong and I can get away with the -i in place sed feature. If that is the case can someone confirm that I will not disrupt the file getting written too?

thanks,
jaysunn

Hi,
based on the hypothesis that filenames structure is always the same:
nnn.nnn.nnn.nnn_nnnnn_yyyymmdd.xxxx
or
nnn.nnn.nnn.nnn_nnnnn_yyyymmdd.xxxx.idx
you may obtain replacement of the <_yyyymmdd.> substring with a new value with instructions like the following:

Code:

dateYest=$(date -d yesterday +%Y%m%d)

for i in `ls /dir/*` 
do

   newfilename=`echo $i | awk -v newdate="$dateYest" { sub(/_[0-9]*\./ , "_" newdate "." ) ; print }`
   mv $i $newfilename 

done

The result will be that the each file in the loop is renamed with the value of variable $newfilename.
At this time I didn't have a chance of testing myself, so do some tests on copies of the original files.

see ya
fra