Shell Programming and Scripting

I have many files which contain about two million lines.
Now I want to use sed to delete the 9th line and add a new
line behind the 8th line. I use the command as follows:

Code:

for((i=1;i<100;i++));
do
  echo $i;
  sed -i '9d' $i.dat;
  sed -i '8a this is a new line' $i.dat;
done

But it is very slow for every file.
Does anybody have a better way to do that?
thank you.

You are reading, overwriting, and deleting the entire file each loop. ...Twice.

Unfortunately if the file is huge, this may never be fast, because you cannot really avoid having to write all the old data. Files do not have an 'insert' or 'delete' operation for data in the middle of them, all the data after the change must be rewritten in the new spot.

This avoids needing to alter the file twice by just printing an alternate line for line 9 instead of deleting then inserting.

Code:

awk 'NR==9 { print "replacement line" ; next } 1' inputfile > outputfile

You might even be able to do it all in one command:

Code:

awk 'FNR==9 { print "Replacement line" > FILENAME".new" ; next } { print > FILENAME".new" }
        (F != FILENAME) { if(F) close(F".new"); F=FILENAME }' *.dat

for FILE in *.new
do
        echo mv "${FILE}" ${FILE%.new}"
done

Remove 'echo' once you've tested and are sure it does what you want.

Use nawk on solaris.

Thanks