Shell Programming and Scripting

Robotronic

Your original awk works just great:

awk -F'"' '
NR==1 {
OFS=FS;
split(FILENAME, x, ".");
prefix=toupper(x[1]"C_");
}
/[Cc][Oo][Uu][Nn][Tt][Yy]/ {
$6=prefix $4;
}
{
print
}' FILENAME_GOES_HERE

My mistake was not realizing that the filename to be processed goes at the bottom of the script.
I thought it was the output from the print command.

The problem now is, how do I pipe several thousand files through the awk?

As you can see I've tried several ways to no avail.

Again, all your help is greatly appreciated.

The loop and the backticks look wrong. Try this variation:

Code:

for i in *.xml; do
  awk -F'"' '
    NR==1 {
      OFS=FS;
      split(FILENAME, x, ".");
      prefix=toupper(x[1]"C_");
  }
  /[Cc][Oo][Uu][Nn][Tt][Yy]/ {
    $6=prefix $4;
  }
  {
    print
  }' "$i"
done >temp

You were overwriting the temp file on each iteration; my guess is that it simply didn't print anything on the last round, so you ended up with an empty result file. Putting the redirection after the done should hopefully work better.

Actually you probably don't even need the loop; try awk '... script goes here ...' *.xml >temp

era is right. If you try this script you don't even need to redirect the output to a temp file: awk does all the work for you:

Code:

rm *.tmp

awk -F'"' '
   FNR==1 {
      OFS=FS;
      split(FILENAME, x, ".");
      prefix=toupper(x[1]) "C_";
      tmpfile=FILENAME ".tmp";
   }
   /<sub_unit / {
      $6=prefix $4;
   }
   {
      print >> tmpfile;
   }
' *.xml

This creates a "*.tmp" file for each input file.

Thanks era,

will give it a try and let you know.

Robotronic

Looks good.

Will give it a whirl and let you know.

Oh, by the way, since this seem no more a Cygwin problem, I may had encountered this issue with Windows Services for UNIX, not Cygwin

Robotronic

I just copy & pasted it into a file and tried to run in under Cygwin
and on my test files here at home it worked perfectly.

Thanks again for all your efforts, I really appreciate the assistance.