Need a little help with using script output text

[legolas_w]

Hi,

I need to use an script from within a script I am writing. the script which I need to invoke as part of my script returns the following text

=========
books=       123
------------
magazines=        628
------------
multimedia=   27
=========

How can I parse this output and put each one of the output values into a separate variable named books, magazines...?

Thank you.

[spynappels]

Can you give a little more info on the scripting language, and whether the output from the first scrip is in a file or just as a stdout stream.

This isn't a homework question, is it?

[papibe]

Hi legolas_w.

What language are you using? Is it a awk script? Bash, Python?

Regards.

[legolas_w]

Sorry, I forgot that we can do the script in many languages. I am using bash (shell scripting). So I assume that awk, sed and perl can be used inside my script to check the output. The output is in stdout stream (screen and not a file).

No it is not homework. I am writing my first large script and some places I can figure out on my own (where we need sed, perl or awk to get the part done.)

Thanks.

[papibe]

You can get the 2nd column of line that start with a known pattern using awk:

awk '/^books=/{print $2}' input.txt

Where:

/^book=/ selects only the lines that start with 'book='
print $2 prints the 2nd column.

The result will be:

123

In order to assign it to a variable in bash you use the $() expression:

books=$(awk '/^books=/{print $2}' input.txt)

I hope that helps, and points you in the right direction.
Regards.

[legolas_w]

Thank you, superb. I fixed the script and can read all variables using the same technique.

am I correct that awk assumes the space between the texts as the field separator by default?

[papibe]

am I correct that awk assumes the space between the texts as the field separator by default?

Indeed.

You can change it easily using awk's -F option. For instance, if you want the '=' to be your separator:

awk -F'='  '/^books=/{print $2}' input.txt

Note that this would put the spaces into the 2nd column now:

       123

(that is a string that stars with several spaces and trails the value '123').

Regards.

[matt_symes]

Hi

There are a whole load of ways to do this and papibe has shown you an excellent example using awk.

I just thought i would show another method using bash parameter expansion.

This is only a very basic script and could be made far more robust. The important lines are highlighted in bold.

#!/bin/bash

book_num=
mag_num=
media_num=

while read line;
do
        # Paramater expansion
        tmp_left=${line%=*};
        tmp_right=${line##* };

        # Match up and save .
        case "$tmp_left" in
                books)
                book_num="$tmp_right";
        ;;
                magazines)
                mag_num="$tmp_right";
        ;;
                multimedia)
                media_num="$tmp_right";
        ;;
                *)
                # ignore ??
                ;;
        esac

done < "$1";

echo "Number of books = $book_num";
echo "Number of magazine = $mag_num";
echo "Number of multimedia = $media_num";

matthew@matthew-Aspire-7540 ~ % cat t
=========
books=       123
------------
magazines=        628
------------
multimedia=   27
=========
matthew@matthew-Aspire-7540 ~

matthew@matthew-Aspire-7540 ~ % ./tt.sh t 
Number of books = 123
Number of magazine = 628
Number of multimedia = 27
matthew@matthew-Aspire-7540 ~ %

For a basic introduction to the commands

http://mywiki.wooledge.org/BashGuide/Parameters

Kind regards