Awk - moving backwards

[nashcom]

Hi

Is it possible to skip backwards a number of records when using awk? For example, if I'm reading each line in an input file and reach a line which begins with certain text I want to skip backwards a couple of lines in the input file and carry out some extra processing.

I may need to revisit the 'logic' in my script, but I thought I'd ask this question anyhow!

Thanks.

 

[feherke]

Hi

Nope.

Whatever you need to process later, you have to store it in variable.

( Or use [tt]tac[/tt] before and after [tt]awk[/tt]... )

Feherke.

http://free.rootshell.be/~feherke/

 

[nashcom]

Thanks for confirming that. I'll tackle the script in a different way!

 

[Annihilannic]

Here is an example of something like that. Every time it finds "four" it prints any records preceding that up to the previous "two" by accumulating them in an array.

awk '
        [green]/two/[/green] { i=0 }
        { a[++i]=[blue]$0[/blue] }
        [green]/four/[/green] { [olive]for[/olive] (j=1;j<=i;j++) [b]print[/b] a[j] ; i=0 }
' << EOF
one
two
three
four
five
one
two
three
two
three
two
three
four
three
four
five

Annihilannic
[small]tgmlify - code syntax highlighting for your tek-tips posts[/small]