Can I shorten/improve this script(grep/cut/touch/mv) 2

[baggetta]

I have this script below that allows me to rename a file based on the word I find in a file. Is there an easier way of doing this? I will be executing this script remotely from another server after it ftp's the file to another server.

#/bin/ksh
grep "*IUNB" t.edw > r.out
cut -c 6-10 r.out > r.in
touch a-riv`cat r.in`.edw
mv t.edw a-riv`cat r.in`.edw

t.edw = source file
r.out = stores the grep line
r.in = stores the name I was looking for
a-riv`cat r.in`.edw will be the new name of the file keeping the original contents of t.edw

 

[SamBones]

Well, if your string you are looking for in the [tt]grep[/tt] appears more than once in [tt]t.edw[/tt], then these commands won't work. If it does appear only once, then this should work...

    #/bin/ksh

    mv t.edw a-riv$(grep "*IUNB" t.edw | cut -c 6-10).edw

The touch isn't really needed since you are moving [tt]t.edw[/tt] to that name in the next step.

The only problem you might have is if there is a character at position 6 to 10 in that line that is invalid for a filename. Such as a space or an asterisk (*).

Hope this helps.

 

[baggetta]

Sweet...thanks this is much nicer...
question, what does the "$" do before the (grep...)

 

[Annihilannic]

The $(grep "*IUNB" t.edw | cut -c 6-10) will be replaced with the output of those two commands.

Try this and you'll get an idea of how it works:

[tt]echo The date is $(date)[/tt]

Annihilannic.

 

[SamBones]

In the Korn shell, using [tt]$(somecommand)[/tt] is the same as [tt]`somecommand`[/tt]. It's a Korn shell equivalent of back ticks.

Hope this helps.

 

[baggetta]

One more problem, how can I cut ahead and then back from the output of grep, can it be done with 1 command?

This is the output of my grep:
*IUNBCHERA ZZ FASJRIV Z
Z UNOA22003090911521002 '+:

Now I need to cut -c 49-51 | cut -c 6-10 from this grep.
Output should now be r-RIVCHERA.edw

 

[PHV]

Try something like this:

mv t.edw r-$(awk '/^\*IUNB/{
printf "%s%s",substr($0,49,3),substr($0,6,5);exit
}' t.edw).edw

Hope This Help
PH.

 

[vgersh99]

nawk '/IUNB/ { printf("%s%s\n", substr($0, 49,51-49+1), substr($0, 6,10-6+1) )}' t.edw

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[baggetta]

When I get the new file name written to disk, how can set an environment variable to use this new file in an ftp command that uses an existing file name of &quot;EDI_JOB_NUMBER&quot; which I don't want to change. Here's part of my script:

mv *${EDI_JOB_NUMBER}* a-$(grep &quot;DESADV&quot; *${EDI_JOB_NUMBER}* | cut -c 6-14).edw
#The command above will rename the EDI_JOB_NUMBER file to this format &quot;a-SITEVENDOR.edw&quot;
#Copy new file name to variable EDI_JOB_NUMBER
#FTP the new file
#
echo open ${SYSTEM} >> ${FTPFILE}
echo user ${USER} ${PASSWORD} >> ${FTPFILE}
echo type ascii >> ${FTPFILE}
echo cd ${PATH} >> ${FTPFILE}
echo mput *${EDI_JOB_NUMBER}* >> ${FTPFILE}

 

[PHV]

IMHO your script can't do what you want.
Do a

 man mv

to see what I mean.
> mv *${EDI_JOB_NUMBER}* a-$(grep &quot;DESADV&quot; *${EDI_JOB_NUMBER}* | cut -c 6-14).edw
The wild chars * in the above command will put your script in trouble.
Try something like this:

echo open ${SYSTEM} >> ${FTPFILE}
echo user ${USER} ${PASSWORD} >> ${FTPFILE}
echo type ascii >> ${FTPFILE} 
echo cd ${PATH} >> ${FTPFILE} 
#         ^^^^ Are you sure ?
for old in $(ls *${EDI_JOB_NUMBER}*); do
  new=a-$(grep "DESADV" $old | cut -c 6-14).edw
  mv $old $new
  echo put $new >> ${FTPFILE}
done

Hope This Help
PH.

 

[baggetta]

PH, I'm getting a Variable syntax. Maybe I should have mentioned this script is being executed in a c shell env.
So if I were try this under straight AIX using ./file, it shouldn't work.
r

 

[Annihilannic]

The first line of your script should read #!/bin/ksh not #/bin/ksh.

If you correct that it will force your script to be interpreted by the Korn shell and the variable syntax error should go away.

Annihilannic.

 

[SamBones]

Good eye Annihilannic! I'll give you my star because I missed that on the first reply!

You don't really need the temporary file $FTPFILE for all the commands. You could just do this...

#!/bin/ksh

# Need to set SYSTEM, USER, PASSWORD, REMOTEPATH
# and EDI_JOB_NUMBER before the rest will work

( print user ${USER} ${PASSWORD}
print ascii
print cd ${REMOTEPATH}
for OLDNAME in $(ls *${EDI_JOB_NUMBER}*)
do
    NEWNAME=a-$(grep "DESADV" ${OLDNAME} | cut -c 6-14).edw
    mv ${OLDNAME} ${NEWNAME}
    print put ${NEWNAME}
done
) | ftp -n ${SYSTEM}

Just enclose the whole thing in parenthesis to make it a subprocess, and pipe it's output to [tt]ftp[/tt].

Hope this helps.

 

[baggetta]

A bit more help. I got the for loop to work as I need it, but is it possible to make the new file name into a variable name so I can ftp the variable name? Since I don't know what these new files will be called, and I don't want to ftp each individual file one by one.

foreach i (`ls ${dd_VLT_IN}*${EDI_JOB_NUMBER}.*`)
setenv rnew $i
echo &quot;EDI SNF File: ${rnew}&quot;
grep &quot;DESADV&quot; ${rnew} | cut -c 6-14 > r_temp
cat r_temp
*** cp ${rnew} a-`cat r_temp`.${EDI_JOB_NUMBER}.${ext}
rm a-`cat r_temp`.${EDI_JOB_NUMBER}.${ext}
rm r_temp
end

*** In this cp command, can the new file become a variable that I can use to ftp? Or do I have to ftp each file 1 by 1 since I know what the r_temp file will be?

 

[PHV]

I'm not confident with C-shell syntax, as I'm haird that csh was broken an many flavors of *nix...
This is my guess:

foreach i (`ls ${dd_VLT_IN}*${EDI_JOB_NUMBER}.*`)
 setenv rnew $i
 echo "EDI SNF File:      ${rnew}"
 setenv anew a-`grep "DESADV" ${rnew} | cut -c 6-14`.${EDI_JOB_NUMBER}.${ext}
 echo " DESADV File:      ${anew}"
end

Anyway, I don't understand why you insist to mput files instead to do multiple put in the same ftp-command file.
If you want to remove the $(anew) files after ftp-ing them, you can store the names within the loop in an another variable, or sed-ing the ftp-command file to retrieve the put statements.


Hope This Help
PH.

 

[baggetta]

PHV, sorry but that did work. I guess you cannot do a setenv statement followed by the &quot;a-grep command&quot;. Came back with Ambigious command. So me being the trail and error person, tried &quot;sentevn anew $a-grep...came back with underfined variable.
Thanks for the help anyways. My script works perfectly when it receives one file. My other problem now is I'm testing when 2 files are received, the script concatenates them before creating 1 file. So my r_temp will have 2 lines in it:
RIV-ABCD
RIV-EFGH
Do you know a command to extract only the first line and save it to file or variable? I figure if I can do this, then were there are multiple files in my r_temp, I can go down the list and pick up the new name to save to my new file.

 

[PHV]

I guess you forgot the 2 ` in the setenv command, or the

 cut -c 6-14

returns trailing space(s).
Yet another guess:

foreach i (`ls ${dd_VLT_IN}*${EDI_JOB_NUMBER}.*`)
 setenv rnew $i
 echo "EDI SNF File:      ${rnew}"
 set x=`grep "DESADV" ${rnew} | cut -c 6-14`
 setenv anew a-${x}.${EDI_JOB_NUMBER}.${ext}
 echo " DESADV File:      ${anew}"
end

Hope This Help
PH.