[SOLVED] Need to remove fist part of files and ensure the new name starts with SCAL

I'm currently trying to rename a series of files I have downloaded using mget. The files are coming from a mainframe so they have dataset qualification names. When I perform the mget the files come down with single quotes surrounding them. I'm running the following script to take those quotes off and was trying to use the same process to remove part of the names. Problem is the part I identify to remove goes away, but so does the first 3 characters of the last name.
The file names are 'ADCDMST.ENG.SCAL147D' 'ADCDMST.ENG.SCAL14PE', and 'ADCDMST.ENG.SCAL14WE'.

I want these files to end up with SCAL147, SCAL14PE, SCAL14WE.
Here is the code I'm currently using to rename.
for filename in `ls *SCAL*`
do
echo $filename
new_filename=`echo $filename| tr "'" " "`
new_filename=`echo $new_filename| tr -d "ADCDMST.ENG."`
echo $new_filename
# mv $filename $new_filename
done

echo "Files Received"

Using this code the files always end up like L147, L14PE, L14WE:
./name.sh
'ADCDMST.ENG.SCAL147D'
L147
'ADCDMST.ENG.SCAL14PE'
L14P
'ADCDMST.ENG.SCAL14WE'
L14W
Files Received


Any assistance would be appreciated. Thanks.

Originally Posted by mfarch99 I'm currently trying to rename a series of files I have downloaded using mget. The files are coming from a mainframe so they have dataset qualification names. When I perform the mget the files come down with single quotes surrounding them. I'm running the following script to take those quotes off and was trying to use the same process to remove part of the names. Problem is the part I identify to remove goes away, but so does the first 3 characters of the last name. The file names are 'ADCDMST.ENG.SCAL147D' 'ADCDMST.ENG.SCAL14PE', and 'ADCDMST.ENG.SCAL14WE'. I want these files to end up with SCAL147, SCAL14PE, SCAL14WE. Here is the code I'm currently using to rename. for filename in `ls *SCAL*` do echo $filename new_filename=`echo $filename| tr "'" " "` new_filename=`echo $new_filename| tr -d "ADCDMST.ENG."` echo $new_filename # mv $filename $new_filename done echo "Files Received" Using this code the files always end up like L147, L14PE, L14WE: ./name.sh 'ADCDMST.ENG.SCAL147D' L147 'ADCDMST.ENG.SCAL14PE' L14P 'ADCDMST.ENG.SCAL14WE' L14W Files Received Any assistance would be appreciated. Thanks.

This thread:
http://www.linuxquestions.org/questi...me-4175513716/

...has a little shell one-liner you could use to do exactly what you're after. Modify the sed parameters, so instead of removing parens, remove the "ADCDMST.ENG."

Note: don't use an extra "ls" in bash scripts. Bash already has good pathname expansion that doesn't suffer the quirks that ls adds. For example, ls is an alias for many users, so you should at least always use /bin/ls to get the real one.

shopt -s nullglob
for  filename in *SCAL*

Tried addig the following to the script but still receive same results. Truncates the last part of the file name.
shopt -s nullglob
for filename in *SCAL*


Also tried to use the one liner and that did not work either. I replaced the code with the following and receive an the error message below; for i in *\(*\)*; do j=`echo $i |sed -e 's/["ADCDMST.ENG"]//g'`; mv "$i" "$j"; done

Error message:
mv: cannot stat `*(*)*': No such file or directory

Originally Posted by mfarch99 Also tried to use the one liner and that did not work either. I replaced the code with the following and receive an the error message below; for i in *\(*\)*; do j=`echo $i |sed -e 's/["ADCDMST.ENG"]//g'`; mv "$i" "$j"; done Error message: mv: cannot stat `*(*)*': No such file or directory

Right, because you didn't modify what was getting found, and have an incorrect sed. The sed doesn't need the brackets or the double-quotes...brackets indicate a regular expression, and the double-quotes aren't escaped. The "for i in..." needs to be the *SCAL*, since that is the pattern it will look for..right now, it's looking for file names with parens in them.

Re-read the other thread, and look at the explanation of what that one-liner is doing...you will have to modify it to make it do what YOU want.

Thanks TB0ne. Looks like I have it.. Sorry for the confusion...

Taking

'ADCDMST.ENG.SCAL147D'

for an example.

Once you've removed the single quotes so that fn = ADCDMST.ENG.SCAL147D, try this

newfn=$(echo $fn | cut -d'.' -f 3)

mv $newfn $fn

thanks

for mainframe i sometimes log in via ftp (since these dinosaur machines dont have ssh/scp).
if i do cd .. then get dsn.name without the single-quotes the downloaded file wont have single-quotes.

alternate:

[schneidz@hyper ~]$ echo first,second,third,fourth | awk -F , '{print $3}'
third

Actually, this can easily be done right within bash, without tr/sed/awk or whatnot:

1. The "classic" way, if you have an older bash:

#!/bin/bash

DIR="$1"
test -z "$DIR" && DIR="."

for i in "$DIR"/*SCAL* ; do
        NEWNAME=${i##*.}         # Strip everything up to last dot
        NEWNAME=${NEWNAME%\'}    # Strip trailing '
        mv "$i" "$DIR/$NEWNAME"
done

2. Or with regular expressions in non-ancient bash:

#!/bin/bash

DIR="$1"
test -z "$DIR" && DIR="."

REGEXP=".*(SCAL.*)'"

for i in "$DIR"/*SCAL* ; do
        if [[ "$i" =~ $REGEXP ]] ; then
                mv "$i" "$DIR/${BASH_REMATCH[1]}"
        else
                echo "$i doesn't match"
        fi
done

This works great. Thanks..

Originally Posted by rupertwh Actually, this can easily be done right within bash, without tr/sed/awk or whatnot: 1. The "classic" way, if you have an older bash: #!/bin/bash DIR="$1" test -z "$DIR" && DIR="." for i in "$DIR"/*SCAL* ; do NEWNAME=${i##*.} # Strip everything up to last dot NEWNAME=${NEWNAME%\'} # Strip trailing ' mv "$i" "$DIR/$NEWNAME" done 2. Or with regular expressions in non-ancient bash: #!/bin/bash DIR="$1" test -z "$DIR" && DIR="." REGEXP=".*(SCAL.*)'" for i in "$DIR"/*SCAL* ; do if [[ "$i" =~ $REGEXP ]] ; then mv "$i" "$DIR/${BASH_REMATCH[1]}" else echo "$i doesn't match" fi done