CODEWITHSUNDEEP
linuxbashshellawkgrep
Darrington Darque
Darrington Darque

Reputation: 21

Take string output from command, and search for that string in a file?

I am wondering how to use the output from a command as the pattern argument of grep? In particular, I have this command I've cooked up (with help from another question):

grep PATTERN file.txt | awk 'NR == 1 {line = $0; min = $5} NR > 1 && $5 < min {line = $0; min = $5} END{print $5}'

that will search file.txt for the line whose 5th column has the lowest value, then print that value. However, I want to then redirect that into grep again, to search file.txt again for that line, so that I may print at least 2 lines above, and 5 lines below (though more isnt' a big deal).

I've looked around, and I know that piping output into grep means you search that output for the argument you give to grep - but I am not sure how to do it the other way around?

Any help would be greatly appreciated. Thanks!

Upvotes: 2

Views: 217

Answers (3)

Jotne
Jotne

Reputation: 41460

Why not do all in one awk

awk '/PATTERN/ && (!min || $5 < min) {min=$5;c=NR} {a[NR]=$0} END{for (i=c-2;i<=c+5;i++) print a[i]}' file

This find the line number where filed 5 has lowest value and store it in c
It also store all lines in array a with line number as index.
Then in end section it prints out from array a 2 lines before and 5 lines after, including the hit line.

Upvotes: 0

Darrington Darque
Darrington Darque

Reputation: 21

Never mind, I figured it out. I can redirect the output to a file, then grep using the -f option to search for the strings in that file. Thanks anyway! Hope this helps if you are reading this with the same problem as me...

Upvotes: 0

merlin2011
merlin2011

Reputation: 75575

If you want to use the output of the command as a command line argment to grep, you can use a subshell with $( ...) syntax.

grep $(grep PATTERN file.txt | awk 'NR == 1 {line = $0; min = $5} NR > 1 && $5 < min {line = $0; min = $5} END{print $5}') Haystack

To get 2 lines of context above and 5 lines below, you can use the -A and -B flags.

grep -B2 -A5 $(grep PATTERN file.txt | awk 'NR == 1 {line = $0; min = $5} NR > 1 && $5 < min {line = $0; min = $5} END{print $5}') Haystack

Upvotes: 1

Related Questions