Search code examples
bash

Check next line in bash

I have a text like:

set 1 recovery 2 2
set 2 recovery 2 2
% ERROR Unknown error
set 3 recovery 2 2
set 4 recovery 2 2
% ERROR Unknown error

What I need is an output like:

set 1 recovery 2 2
set 2 recovery 2 2 ==> % ERROR Unknown error
set 3 recovery 2 2
set 4 recovery 2 2 ==> % ERROR Unknown error

I tried with awk and sed, but I'm failing. What can I try next?

Solution

  • Assumptions:

    • there's at least one non-ERROR line before each ERROR line
    • no 2 consecutive lines are ERROR lines

    Sample data:

    $ cat check.dat
set 1 recovery 2 2
set 2 recovery 2 2
% ERROR Unknown error
set 3 recovery 2 2
set 4 recovery 2 2
% ERROR Unknown error

    One awk solution where we rely on printf to leave us on the current line until we explicitly tell it to print a newline (\n):

    awk '
/ERROR/ { printf " ==> %s",$0 ; next }
        { printf "%s%s",nl,$0 ; nl="\n" }
END     { printf "\n" }
' check.dat

    Where:

    • /ERROR/ ... - if current line includes the string ERROR then append ==> and the current line to the output; then skip rest of processing and go to the next line of input
    • { printf "%s%s",nl,$0 ; nl="\n" } - for lines w/out ERROR we print the line but first we print the nl variable; initially nl="" so for the first line of input we'll print to the first line of output; for successive non-ERROR lines we want nl="\n" so we can move the cursor off the end of the previous line and down to a new line before we print our non-ERROR line of text
    • END ... - print one last new line characters (\n)

    Results of running the above:

    set 1 recovery 2 2
set 2 recovery 2 2 ==> % ERROR Unknown error
set 3 recovery 2 2
set 4 recovery 2 2 ==> % ERROR Unknown error