Grep search from end of file

Question

I have large log file 300-500 megs and need to do searches in it. I am currently doing this using grep in php code, but I want to further optimize the code. Data I need is always within last 10000 lines in the file so I want the search to start from the end and move forward until pattern is found and stop.

I found some info that tac can do this but my performance tests show using tac is 4 times slower. Currently my search takes about 0.13 sec to complete with tac added on it takes 0.50. I feel like I am missing something, I would appreciate help.

My expectation is that if I get file read from the end properly this will increase the speed of the process considerably.

here is the code and it includes the timer

<?php
exec('grep -r "search var" file.log -b  | head -1', $result);
echo $result[0];

$time = microtime(true) - $_SERVER["REQUEST_TIME_FLOAT"];
echo "<br>".$time; 
?>

Answer 1

You can avoid the -r flag with grep . And you only want to search within the last 10000 lines right? So directly tail it and grep the pattern.

tail file.log -n 10000 | grep "search var"