Hi,

i have a large text file that contains several portions of text. Here is an example:
"Log 1:

mpla
mpla ok ok ,pla mpla mpla
mpla
mpla

Log2:
mpla2 mpla2 mpla2 mpla2
mpla2 mpla2
mpla2 mpla2

Log3:
mpla? mpla?

Log4:
&&& mpla
"

I want to split the above file into 4 different files, which will contain the appropriate log entries. That is:
file1 => Log1:
mpla
mpla ok ok ,pla mpla mpla
mpla
mpla

file2 => Log2:
mpla2 mpla2 mpla2 mpla2
mpla2 mpla2
mpla2 mpla2

etc.

Any template available using awk, sed or something alike?

Thanks

With perl it's quite easy...
Look and learn
put this in a file, make it executable and run.

Enjoy
nukkel

Very good script.
Thanks

How can I check if a file with the same name (e.g. log1) exists in the output directory in order not to erase it?

Something along the lines of

Thanks. I needed the stat function anyway to test the modification time of the file.

Another thing is that I openned a file for logging and I write something to it but nothing is written until I close the file handle. Is there a flush equivalent in Perl?

I have also problem displaying the current time as string using localtime function, let's say "Fri 21 May 2004 11:08:45". Any suggestion?

Thanks again

Do you know how can I check if a file is modified without polling? I mean now i check in endless loop the file and check every 30 secs the modification of the file.

Thanks

Try if this works for auto-flushing the output file:

use IO::Handle;
open(FILE, ">$file");
FILE->autoflush(1);

For the current localtime try
print scalar localtime();

for the polling... you can use select() at the beginning of the main loop so the loop only gets executed when the file changes. But the downside is your program can't do anything else in the meantime. I forgot the exact syntax though... Think it's a bit like the posix select() used in C (man 2 select)

Have fun programming

Ok. here is my code after your suggestions:

The strange things are the following:
1. START is never printed
2. $currentFileTimeSize never printed
3. "if block" executed only once, even though the file is modified

Any suggestions?

1 & 2: It's the more or less same problem you had on the output file: the buffer for stdout only gets flushed when a newline is written... So print "START\n", with a \n, is better.

3: when you use (stat(...))[7] you should delete the 'use File::stat' line, because the File::stat package overrides the stat() function with a little more user-friendly one where you can say '$filesize = (stat($file))->size'
So you don't have to look up which array member is which number

So like you did it you should remove the 'use File::stat' and it should work

I left the use File::stat and used your aproach (more readble).

Thanks a lot

No problem... See you around

Hello,

I am reading your previous post on the linuxquestions.org regarding split large file into severel small files using perl. I got the exact same problem as you had in the post. I tried your perl script (I just modified the input and output filename). I got the perl compile error
Can't locate object method "autoflush" via package "IO::Handle" (perhaps you forgot to load "IO::Handle"?) at wbsplit.pl line 15.

Here is my script

#!/usr/bin/perl
#
use strict;

my ($line, $nr);

my $thebigfile = "/home/oracle/projects/test/wbreports.txt"; # input file location
my $logfile = "/home/oracle/projects/test/newwb"; # output files basename

my ($previousFileTimeSize, $currentFileTimeSize);
$previousFileTimeSize = 1;

print "START\n";
open(LOGFILE, ">$logfile");
LOGFILE->autoflush(1);
while (1) {
$currentFileTimeSize = (stat($thebigfile))[7]; # size
print $currentFileTimeSize;
if ($currentFileTimeSize != $previousFileTimeSize) {
print LOGFILE scalar localtime;
print LOGFILE ": sent-mail MODIFIED\n";
$previousFileTimeSize = $currentFileTimeSize;
} else {
print LOGFILE scalar localtime;
print LOGFILE ": sent-mail no modification\n";
}
sleep 30;
}
close LOGFILE;

Do you have any idea what went wrong in the code? I would appreciate your time and help.

Thanks
Michael

I think you'll need to put "use IO::Handle" at the beginning of the script, before the "autoflush" function can be used. Let me know if that works out for you.

Best regards,
nukkel

Hi Nukkel,

Thank you so much for your reply. I have put the "use IO::Handle;" in the top of the program. After I executed the program, it did print out the START. Then it just hangs over there for ever. I guess we just open the IO and didn't really write the lines to the file.
Here is the resluts of executing the program
oracle@dbsdata.nrtc.org:twsb$erl wbsplit.pl
START

Any ideas?

Thanks
Michael

just replace log text with some string, just log

u can simply use

awk '/log/{n++}{print > f n}' f=destination file source file

But the problem is that u can only have atmost 10 files, even i am looking for some other one line command which can make more than 10 files.