Whats the proper way to get last line of a file...

Dev Shed Forums Sponsor:

May 23rd, 2006, 02:27 AM

papajohns

perl virgin

Join Date: Oct 2003

Location: Seattle, WA

Posts: 454

Time spent in forums: 2 Days 13 h 23 m 58 sec
Reputation Power: 12

Whats the proper way to get last line of a file...

whats the "proper way" to do it if you are building as a part of a .pm to be used by other perl files...

a simple
$variable = `tail -f file`;

or open file, read it , then get the last line...

or a system call...

or is there a built in perl call for it?

which is it?

May 23rd, 2006, 03:23 AM

fuzzybunny

C Neophyte.

Join Date: Feb 2004

Location: Melbourne Australia

Posts: 398

Time spent in forums: 3 Days 42 m 6 sec
Reputation Power: 43

Not very clean, as it doesn't get the last line... it goes in the number of bytes from the end. Number 2 indicates end of file, 9 bytes back.

perl Code:

Original - perl Code
seek( FILE, -9, 2 ); my @line = readline FILE;

seek( FILE, -9, 2 );
my @line = readline FILE;

May 23rd, 2006, 03:46 AM

KevinADC

Contributing User

Join Date: Oct 2004

Location: Sunny Southern California

Posts: 2,687

Time spent in forums: 3 Weeks 4 Days 12 h 58 m 13 sec
Reputation Power: 859

You can use Tie::File or maybe File::Bidirectional or File::ReadBackwards

May 23rd, 2006, 04:39 AM

Axweildr

'fie' on me, allege-dly

Dev Shed God 16th Plane (12500 - 12999 posts)

Join Date: Mar 2003

Location: in da kitchen ...

Posts: 12,668

Folding Points: 161075 Folding Title: Super Ultimate Folder - Level 1

Time spent in forums: 4 Months 2 Weeks 3 h 7 m 53 sec
Reputation Power: 6120

Send a message via Google Talk to Axweildr

using seek, you could pull in a big chunk of data into an array, reverse it, join it into a string, use index to find the location of the first "\n", substr to extract the last line in reverse, split on '' into an array, then reverse and join again ...

__________________
--Ax
without exception, there is no rule ...
Heavy Haulage Ireland
Targeted Advertising Cookie Optout (TACO) extension for Firefox

The great thing about Object Oriented code is that it can make small, simple problems look like large, complex ones

09 F9 11 02
9D 74 E3 5B
D8 41 56 C5
63 56 88 C0
Some people, when confronted with a problem, think "I know, I'll use regular expressions." Now they have two problems. -- Jamie Zawinski
Detavil - the devil is in the detail, allegedly, and I use the term advisedly, allegedly ... oh, no, wait I did ...

May 23rd, 2006, 10:04 AM

replax

Contributing User

Dev Shed Intermediate (1500 - 1999 posts)

Join Date: Aug 2004

Posts: 1,795

Time spent in forums: 2 Weeks 5 Days 15 h 23 m 11 sec
Reputation Power: 82

Quote:

Originally Posted by papajohns

a simple
$variable = `tail -f file`;
...
or a system call...

`tail -f file` is a system call. backticks are the system operator. It is just as if you were ssh'ed into the sytem and you typed that at the command line. You are calling the executable tail and feeding it some command line parameters.
So we are down to 2 methods:
a system call (such as tail)
or reading the file manually, your choice.

May 23rd, 2006, 01:00 PM

Axweildr

'fie' on me, allege-dly

Join Date: Mar 2003

Location: in da kitchen ...

Posts: 12,668

Time spent in forums: 4 Months 2 Weeks 3 h 7 m 53 sec
Reputation Power: 6120

Technically the seek option isn't the same since you can get the length of the file from the stat function and set the file pointer accordingly. Tie::File, I *assume* requires reading the file into an array, because how you can treat the file when Tied, File::ReadBackwards, I'm not sure about, haven't looked at the code.

So there's at least three

May 23rd, 2006, 01:10 PM

replax

Contributing User

Join Date: Aug 2004

Posts: 1,795

Time spent in forums: 2 Weeks 5 Days 15 h 23 m 11 sec
Reputation Power: 82

yes, and there are 5 ways to do anything in PERL. So we just need someone to point out the other 2

May 23rd, 2006, 01:15 PM

ishnid

kill 9, $$;

Dev Shed God 4th Plane (6500 - 6999 posts)

Join Date: Sep 2001

Location: Dublin, Eire

Posts: 6,819

Time spent in forums: 4 Months 1 Week 5 Days 19 h 23 m 31 sec
Reputation Power: 3763

Oh well, since you asked

perl Code:

Original - perl Code
my $last_line; while(<FILE>) { $last_line = $_ if eof; } print $last_line;

 
my $last_line;
while(<FILE>) {
   $last_line = $_ if eof;
}
print $last_line;

Comments on this post

fuzzybunny agrees: Nice... is there a way of starting from EOF to avoid looping through the whole file?

__________________
~ishnid;

Have you tried:
[ search.cpan.org | perldoc | Java API | mysql.com | google ]

Apostrophes are NOT used for possessive pronouns or for noun plurals, including acronyms.

May 29th, 2006, 03:26 PM

papajohns

perl virgin

Join Date: Oct 2003

Location: Seattle, WA

Posts: 454

Time spent in forums: 2 Days 13 h 23 m 58 sec
Reputation Power: 12

Quote:

fuzzybunny agrees: Nice... is there a way of starting from EOF to avoid looping through the whole file?

yah..thats what i want to know..is there a faster way rather than start at the top..

#10

May 29th, 2006, 04:20 PM

Axweildr

'fie' on me, allege-dly

Join Date: Mar 2003

Location: in da kitchen ...

Posts: 12,668

Time spent in forums: 4 Months 2 Weeks 3 h 7 m 53 sec
Reputation Power: 6120

seek as alluded to earlier, would be quicker in this context, any idea how long an average record is?

#11

May 29th, 2006, 05:52 PM

papajohns

perl virgin

Join Date: Oct 2003

Location: Seattle, WA

Posts: 454

Time spent in forums: 2 Days 13 h 23 m 58 sec
Reputation Power: 12

mm..but from the example there it seems seek needs a number of bytes back from the end of the file...

anyway you can use seek, but then only get it up to the last carriage return so you only get the last line as oppose to a certain of bytes and then have to regex it somehow?

#12

May 30th, 2006, 04:33 AM

Axweildr

'fie' on me, allege-dly

Join Date: Mar 2003

Location: in da kitchen ...

Posts: 12,668

Time spent in forums: 4 Months 2 Weeks 3 h 7 m 53 sec
Reputation Power: 6120

Do we know the record length, or any idea of the max possible record length?

set the pointer to file size - max rec length*2 (to be sure)
read for max rec length * 2
split the buffer on \n into an array
and the last populated item in the array is your last line

#13

May 30th, 2006, 06:55 PM

NullTerminator

Registered User

Join Date: May 2006

Posts: 18

Time spent in forums: 4 h 15 m 11 sec
Reputation Power: 0

You want last line and lines are delimited by newlines/

why not
`tail -n 1 file`

Jeb
\0

#14

May 30th, 2006, 08:16 PM

Scorpions4ever

Banned ;)

Dev Shed God 8th Plane (8500 - 8999 posts)

Join Date: Nov 2001

Location: Woodland Hills, Los Angeles County, California, USA

Posts: 8,978

Time spent in forums: 1 Month 2 Weeks 6 Days 22 h 11 m 29 sec
Reputation Power: 3561

The proper PORTABLE way would be to read the file manually. You can either start from the beginning and work your way through all the lines, but the faster way would be to seek to the end and read chunks of the file and use rindex() to find \n.

Any solutions that use the tail command are assuming you're running on a unix or unix-based system (or have cygwin installed). Hence, they are potentially unportable.

Comments on this post

ishnid agrees!

bhagwatn agrees!

__________________
Up the Irons
What Would Jimi Do? Smash amps. Burn guitar. Take the groupies home.
"Death Before Dishonour, my Friends!!" - Bruce D ickinson, Iron Maiden Aug 20, 2005 @ OzzFest
Down with Sharon Osbourne
Diary of a first time dog owner <-- my cousin's blog

#15

January 19th, 2008, 11:21 PM

tomlobato

Registered User

Join Date: Jan 2008

Posts: 1

Time spent in forums: 5 m 14 sec
Reputation Power: 0

After seek to the end of file, it rewind at 1 byte steps until find the begin ("\n") of the last line.

my $pos=-1;
my $char;
my $already_nonblank = 0;
while (seek (LOGFILE,$pos--,2))
{
read LOGFILE,$char,1;
last if ($char eq "\n" and $already_nonblank == 1);
$already_nonblank = 1 if ($char ne "\n");
}
my $last_line = <LOGFILE>;

Tom Lobato
www tinecon com br