Awk command on terminal

Ordinarily, I stick with what is installed with OS X. This awk example uses GNU awk (gawk), and I have installed homebrew (brew) package manager on my OS X El Capitan, and then used brew install gawk to put it into /usr/local/bin.

Although awk on OS X accepts the ENDFILE syntax without error, it does nothing when the end of an individual file is reached. It does work in gawk though.

Here is the awk script that you run in the Terminal, and provide *.txt files on the command line. It will process every line in these files looking for Deet, and when it reaches the end of the given file (ENDFILE), it will print out the last captured DEET, and the current filename. I called the following script deet.awk, and you run it as:

./deet.awk *.txt

#!/usr/local/bin/gawk -f

#

# capture and print the last match for word in its file

function usage()

{

printf "Usage: %s *.txt\n", ENVIRON["_"]

}

BEGIN {

if (ARGC < 2) {

usage()

exit 1

}

}

/Deet/ { match($0, /(Deet)/, z);a = z[1];b=FILENAME }

ENDFILE { printf "%s\t%s\n", a, b }

END { exit }

Output:

odin: ~/awkism$ ./deet.awk *.txt

Deet test10.txt

Deet test25.txt

Deet test30.txt

Deet test45.txt

Hello

You may try something like this:

#!/bin/bash # print filename and query text awk -v q=Deet ' $0 ~ q { a[FILENAME] = $0 } END { for (f in a) { printf "%s\t%s\n", f, q } } ' *.txt

or this:

#!/bin/bash # print filename and the last matching line in the file awk -v q=Deet ' $0 ~ q { a[FILENAME] = $0 } END { for (f in a) { printf "%s\t%s\n", f, a[f] } } ' *.txt

Regards,

H

PS. If you wish, you may create a file named find_last.awk with the following contents

#!/usr/bin/awk -f # # file: # find_last.awk # usage: # ./find_last.awk -v q=Deet *.txt # ./find_last.awk q=Deet *.txt # $0 ~ q { a[FILENAME] = $0 } END { for (f in a) { printf "%s\t%s\n", f, a[f] } }

and use it like this:

#!/bin/bash ./find_last.awk -v q=Deet *.txt

or even this:

#!/bin/bash ./find_last.awk q=Deet *.txt

Good luck,

H

PPS. Changed command name from find.awk to find_last.awk.

Any reason why you need to use AWK?

This can be done with grep and tail:

for f in ~/Desktop/*.txt; do grep -H 'Deet' "$f" | tail -1; done

This will search .txt files on the Desktop

(You can also print the line # in the file with grep -n)

Awk notwithstanding, the requirement was to display the last matched "Deet" word with the filename. If you modify your grep example as follows, it will show the matched, last "Deet" word with the filename. As it stands now, it will display the entire line containing "Deet."

for f in *.txt;do grep -Ho '(Deet)' "$f" | tail -1;done

For some odd reason, -Ho didn't work when I tested it (thought -o was overriding -H), but now it's working (previous typo on my part, I guess)

(I think you need to remove the () in your example)

Lots more options for the OP in man grep (i.e. -i for case insensitive searches)

grep seems the better solution for this problem.

I keep forgetting that in my .bashrc file, I have the following set, so I don't have to explicitly use grep -E, or egrep for extended regular expressions:

export GREP_OPTIONS='--color=auto --extended-regexp'

Still, even when I unset GREP_OPTIONS, and follow that with your syntax, whole lines are printed. One can omit the capture parentheses, keep the -Ho arguments, and the last "Deet" in the line is shown.

In .bash_profiile I use: alias grep='grep —color=auto —line-number —ignore-case'

instead of export GREP_OPTIONS='—color=auto —line-number —ignore-case'

…so when I want to run without options. I just escape grep: \grep

Not sure why OP was searching for the last occurrence of Deet.

Without context or line number, it's the same output as the 1st occurrence of Deet