My field separator changes when using awk
 

Hi,
I have this file:
Helene;/home/home.html;24/02/04 14:16:09;oldcomment
Helene;/home/home.gif;24/02/04 14:16:09;oldcomment

With this code, I want to change the comment (last field of the record), to a new comment specified by the user.

variable="Helene"
awk -F';' ' { if( $1 ~/'${variable}'/ ){
print "Please enter a comment for this record: " >"/dev/tty"
print $0 > "/dev/tty"
getline newcomment < "/dev/tty"
$5=newcomment
print $0 }
} ' myFile > tmpFile

This is the result:
Helene /home/home.html 24/02/04 14:16:09 newcomment1
Helene /home/home.gif 24/02/04 14:16:09 newcomment1

Where is my field separator going?? :eek:

You're declaring the semi-colon as a field separator and only telling it to print fields. So the semi-colon isn't being printed because it's not part of a field. My awk is terrible so I don't know the right way - I think there's an elegant way around this but a stubborn 'I want a semi-colon!' method would be

Something along those lines.

Thanks Digiot. That's the same solution as I came up with my self. It doesn't look very good, but it works!! :) Thanks anyway..

To make it "look better" you should also set an output separator ("OFS"). AFAIK this cannot be done from a command line option, but needs to be set from inside the awk-script. So below the input separator ("FS") is also set inside the script instead of with the -F';' option.