CODEWITHSUNDEEP
awk
dhatfield
dhatfield

Reputation: 3

Using file redirects to input a variable search pattern to awk

I'm attempting to write a small script in bash. The script's purpose is to pull out a search pattern from file1.txt and to print the line number of the matching search from file2.txt. I know the exact place of the pattern that I want in file1.txt, and I can pull that out quite easily with sed and awk e.g.

sed -n 3p file1.txt | awk '{print $4}'

The part that I'm having trouble with is passing that information again to awk to use as a search pattern in file2.txt. Something along the lines of:

awk '/search_pattern/{print NR}' file2.txt

I was able to get this code working in two lines of code by storing the output of the first line as a variable, and passing that variable to awk in the second line, 

myVariable=`sed -n 3p file1.txt | awk '{print $4}'`
awk '/'"$myVariable"'/{print NR}' file2.txt

but this seems "inelegant". I was hoping there was a way to do this in one line of code using file redirects (or something similar?). Any help is greatly appreciated!

Upvotes: 0

Views: 174

Answers (1)

Jonathan Leffler
Jonathan Leffler

Reputation: 754060

You can avoid sed | awk with

awk 'NR==3{print $4; exit 0}' file1.txt

You can do your search with:

search=$(awk 'NR==3{print $4; exit 0}' file1.txt)
awk -v search="$search" '$0 ~ search { print NR }' file2.txt

You could even write that all on one line, but I don't recommend that; clarity is more important than brevity.

In principle, you could use:

awk 'NR==3{search = $4; next} FNR!=NR && $0 ~ search {print NR}' file1.txt file2.txt

This scans file1.txt and finds the search pattern; then it scans file2.txt and finds the lines that match. One line — even moderately clear. There'll be lots of matches if there isn't a column 4 on line 3 of file1.txt.

Upvotes: 2

Related Questions