grep a log file between 2 dates

Shell Programming and Scripting
1

Hi

Currently I can grep a log file with the following command:

$results = `grep -A 2 '^$date.$time.*' $log`;

and the following arguments:

$date = 2007/04/25
$time = 16:07

Log example:

2007/04/25 16:07:12.145701
2007/05/25 14:07:12.145701
2007/05/25 17:07:12.145701
2007/06/25 16:07:12.145701

Therefore running the script ./script 2007/04/25 16:07

will return the 2nd line of the log - 2007/05/25 14:07:12.145701

But how do I do range of dates, by giving a 2 more arguments:

./script 2007/03/20 15:13 2007/08/19 14:31

So therefore returning:

2007/04/25 16:07:12.145701
2007/05/25 14:07:12.145701
2007/05/25 17:07:12.145701

I dont see how I can do that with grep as it is just pattern matching once, do I need a loop which can step through all the days and seconds in between the 2 dates or is there a kind of date range function?

Thanks
Rich

2

non trivial, grep will not cope

3 
awk '$0>=from&&$0<=to' from="2007/03/20 15:13" to="2007/08/19 14:31" infile

Use nawk on Solaris.

4

of course, brilliant

5

im using, is this compatible:
SuSE Linux 9.3 (i586)
VERSION = 9.3

'$0>=from&&$0<=to' is ensuring that the $0 is in between the from and 2 dates specified.

How can I run this in perl?

Im using:
$results = `awk '$0>=from&&$0<=to' from="$date $time" to="$date2 $time2" $log `;

then "if (!$results) {" to find out if $results contains anything? but is not working

error is:

awk: ./s>=from&&./s<=to
awk: ^ syntax error

6

yes is working in the shell (command prompt) but not when running as part of perl script?.

How can I incorporate that into my perl script?

7

Write it in perl (not awk) ...

8

hmm yes......anyway

I dont know how, hence asking here.

9

I understand that :slight_smile:
I don't know perl, so I'm not answerin' your question.
I'm just tryin' to say that if it's a perl script,
you should write it all in perl (as far as I know, it's quite feature-rich and powerful).

P.S. I shoud have noticed -> $results= ...

10 
perl -ne ' print if ( $_ >="2007/03/20 15:13" && $_ <= "2007/08/19 14:31" ) ' filename
11

OK, that also works from the command prompt.

But I am trying to incorporate it into my script so therefore:

  • I need to assign the results of the search to a variable.
  • Do I need to assign to an array or can I assign the search results to 1 variable

$results = `print if ( $_ >="2007/03/20 15:13" && $_ <= "2007/08/19 14:31" ) ` $log;

12 
#!/usr/bin/perl
while( <> )
{
        $results = $results . $_ if ( $_ >="2007/03/20 15:13" && $_ <= "2007/08/19 14:31" )
}
13

oh where has reading in the file gone?

14 
Scriptname filename
15

In the above code, where I am specifying my log file to read from, in your previous codes "filename"?

16

Script.pl

#!/usr/bin/perl
while( <> )
{
        $results = $results . $_ if ( $_ >="2007/03/20 15:13" && $_ <= "2007/08/19 14:31" )
}

Run the script with your filename as

Script.pl your_log_filename
17

ahh ok I get it. Last question now hopefully.

I already have assigned a variable to the name of the log = $log

so instead of the while (<>) which is just reading from the argument, how can I change this for $log?

18 
#!/usr/bin/perl
$log="filename";
open FILE , $log;
while( <FILE> )
{
        $results = $results . $_ if ( $_ >="2007/03/20 15:13" && $_ <= "2007/08/19 14:31" )
}
19

just realised this is not working at all, just gives me back every line and not filtering by the date and time etc

runing:

perl -ne ' print if ( $_ >="2007/05/20 15:13" && $_ <= "2007/08/19 14:31" ) ' filename

2007/05/25 15:51:36.957753
2007/05/25 15:51:36.958117
2007/05/25 15:56:06.670401
2007/05/25 15:56:06.670683
2007/05/25 16:07:12.145404
2007/05/25 16:07:12.145701
2007/04/14 16:07:12.145701

u can see that im getting the last entry too - 2007/04/14 16:07:12.145701

20

the log file actually contains more than just the date and time of course

2007/05/25 15:51:36.957753 dsfsdf dsfdfdf-sdfdfdf Trace name: sfsdfdsf

So shall i just add .* after the time like with regex?