Grep result in a loop and linefeed issue

[RomainPelissier]

Hi,
I try to grep a file (syslog file) and display the result using a loop. Unfortunatly, there is an issue with the code below because of the libefeed/carriage return.
I try to find a way to have each grep result into a single line so I can use it into a the loop:

for res in $(grep 'pam_unix' $LOG_DIR/$server--$YEAR-$MONTH_D-$DAY_D_ZERO.log | grep 'session opened' | grep -v 'closed'); d

                echo "res : $res"

done

Right now, each 'res' results is a single word from the grep result... I can't find a way to catch the end of line and put the the line into the 'res' variable.

Any help will be greatly appreciated!
Thanks

[khafa]

hi,

i think this not a grep problem but rather a for loop problem. for considers the white space as separator. so it takes one word after the other and puts it in res. using while and read should solve the problem

[RomainPelissier]

Yes! it works! :

grep 'pam_unix' $LOG_DIR/$server--$YEAR-$MONTH_D-$DAY_D_ZERO.log | grep 'session opened' | grep -v 'closed' | while read res ; do
echo "res : $res"

done

[khafa]

good news !

[ghostdog74]

grep pipes to standard output by default. Therefore, there's really no need for a while loop to specifically echo the results out. (unless there are further processing on the lines itself)
Secondly, too many unnecessary greps. Unix enables you to use pipes, which is good, but too much of it waste processor times.

awk '/pam_unix/ && /session opened/ && !/closed/'  $LOG_DIR/$server--$YEAR-$MONTH_D-$DAY_D_ZERO.log

[khafa]

grep pipes to standard output by default. Therefore, there's really no need for a while loop to specifically echo the results out. (unless there are further processing on the lines itself)
Secondly, too many unnecessary greps. Unix enables you to use pipes, which is good, but too much of it waste processor times.

awk '/pam_unix/ && /session opened/ && !/closed/'  $LOG_DIR/$server--$YEAR-$MONTH_D-$DAY_D_ZERO.log

i dont think his/her objective is to output .i thought that s/he is gonna use the result for further processing, that is why i suggested while and read. if its just to display then of course you dont need all that.
for the second part i agree with you.

[scm]

Secondly, too many unnecessary greps. Unix enables you to use pipes, which is good, but too much of it waste processor times.

I'd certainly agree with that, but against that you have to consider whether a few greps doing simple pattern matching is actually faster than a single (say awk) command doing a much more complicated one. And I've had at least one of those!