use variable with grep

Question

I am trying to extract values from email header, as the Message ID

The grep do not properly use the previous variable...

Here is my code:

for email in $1/*.eml
do

MSGID=$(cat ${email} | grep -i 'Message-ID:' |awk '{print $2}')
M=$(echo "\""${MSGID}"\"")
SRV="/var/log/maillog"
echo "${M}"

VER=$(grep -i ${M} ${SRV} | tail -n 1)
echo ${VER}
done

I am trying to re-encapsule the initial variable MSGID with M variable; Unfortunately I still could not figure this out.

Initial value (input) looks like this:

Message-ID: <COMPUTER123DDF38c28d5e776e4fbdb82e2e92bd9a4373@computer123ddf>

The MSGID keep only the string as:

<COMPUTER123DDF38c28d5e776e4fbdb82e2e92bd9a4373@computer123ddf>

Then I would like to use this string against a file using grep. Unfortunately the VER command does not trigger anything.

I thought that adding "\""${MSGID}"\"" would/should have keep the initial string untouched.

There are a lot of redundant steps in this script. Please edit your question to provide a sample of your input and the desired output. — Tom Fenech
– Tom Fenech, Commented Mar 2, 2015 at 11:53
Please show some input file together with your expected output. Maybe it is worth not making your current script the starting point, since for example awk may handle it in a way straight-forward way. — fedorqui
– fedorqui, Commented Mar 2, 2015 at 11:54
Probably MSGID=$awk '/^Message-ID:/ {print $2}' file) would be faster. Also, note SRV can be declared outside the for loop, since it is a fixed value. — fedorqui
– fedorqui, Commented Mar 2, 2015 at 12:18

tripleee · Accepted Answer · 2015-03-02 12:34:21Z

2

It's not really clear what you are attempting to accomplish, but just removing the weird stuff simplifies to a script which might help you finish your task.

SRV="/var/log/maillog"
for email in $1/*.eml; do
    M=$(awk 'tolower($0) ~ /message-id:/ {print $2}' "$email")
    echo "$M"
    grep -i "$M" "$SRV" | tail -n 1
done

This has a number of problematic corner cases. The Message-Id extraction should probably be stricter:

M=$(awk 'tolower($1) ~ /^message-id:/ { print $2 } /^$/ { exit(0) }' "$email")

and if the extracted Message-Id contains regex metacharacters, you should escape them, or perhaps use proper grep flags:

grep -iF "$M" "$SRV" | tail -n 1

To be sure, the double quotes in grep "$string" file are not part of the syntax for grep; they are discarded by the shell before grep runs, and are useful for preventing the shell from messing with whitespace and/or shell metacharacters in the value of $string. When you forced in another pair of double quotes, you were grepping for the Message-id inside literal double quotes in the log file, and (of course) not finding any.

edited Mar 2, 2015 at 12:34

answered Mar 2, 2015 at 12:14

tripleee

192k37 gold badges318 silver badges368 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

fedorqui Over a year ago

I don't know the lc() function in awk. To lowercase you probably want to use tolower(). Note you can also say IGNORECASE=1 in the BEGIN{} block.

Arun A S · Accepted Answer · 2015-03-02 12:23:56Z

0

Try this solution

 M="\""$MSGID"\""

edited Mar 2, 2015 at 12:23

Arun A S

7,0704 gold badges33 silver badges54 bronze badges

answered Mar 2, 2015 at 11:54

Peter

476 bronze badges

Comments

fedorqui · Accepted Answer · 2015-03-02 12:24:45Z

0

Thank you all. And sorry.

I have figured that there were a carriage return in my initial value.

I have changed this like this:

MSGID=$(cat ${email} | grep -i 'Message-ID:' |awk '{print $2}' tr -d '\r')

Thank you again.

edited Mar 2, 2015 at 12:24

fedorqui

293k113 gold badges592 silver badges640 bronze badges

answered Mar 2, 2015 at 12:17

user2335924

551 silver badge10 bronze badges

Collectives™ on Stack Overflow

use variable with grep

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related