Manipulate a fixed length file using sed/awk

secret_squirrel

I have a fixed length file that I want to reformat.

Each record consists of 25 chars of 'header' info, then either 1 or 2 subrecords of length 260. So each record looks like this
<..25..><...260...><...260...>\n or
<..25..><...260...>\n

At the moment I have this code to take split each 2 subrecord record into 2 records.
eg
<..25..><...subrec1...><...subrec2..>\n

becomes

<..25..><...subrec1..>\n
<..25..><...subrec2..>\n

This is my prototype code

#create single leg file
cut -c 1-285 ${mqfile} > mq1.txt
#create 2nd leg file
cut -c 1-25,286-546 ${mqfile} > mq2.txt
#cat files
cat mq1.txt mq2.txt > mq1.txt

It works but Im sure there is a better way to this with sed and/or awk which will preferably let me maintain the original record order.

I need a hint in the right direction! Oh and perl etc isnt available to me.

krinDar

secret_squirrel wrote:

I have a fixed length file that I want to reformat.

One method in awk is using the substr function.

if you cat this into a file:

{
       header=substr($0,0,26);
       rec_one=substr($0,26,260);
       rec_two=substr($0,285,260);

      if ( length(rec_one) > 5 ) 
      {
            print header rec_one     
      }

     if ( length(rec_two) > 5 )
     {
             print header rec_two 
     }

}

and run it as follows:
awk -f file.awk record_file.txt

This is pretty kludgey, but it does the job and should be straightforward to use. The strength of awk is that the breaks lines up based on seperators, such as commas, but it is not so handy with fixed length records.

You may need to change some of the values it uses in substr to suit your example.

secret_squirrel

That worked a treat - many thanks. Seems more efficient too.