File Size

[Scott McDade]

Is there an easy way to programmatically read file size either from command line or via script or possible Ansci C funciton?

OS -> 11i

[Dave Olker · Accepted solution]

Hi Scott,

The stat(2) system call is probably the most common means of getting this information programmatically.

Look at the stat(2) man page:

off_t st_size; /* File size (bytes) */


The st_size field contains the size of the file in bytes.

Regards,

Dave

[James R. Ferguson]

Hi Scott:

Perl has access to the 'stat(2)' function. To list the names of files and their size in bytes, you can do:

# perl -le 'die unless @ARGV;print join " = ", $_, (stat($_))[7] for @ARGV' filename

For example:

# perl -le 'die unless @ARGV;print join " = ", $_, (stat($_))[7] for @ARGV' /etc/hosts /etc/services /etc/rc.config.d/netconf

...might return:

/etc/hosts/ = 606
/etc/services = 9946
/etc/rc.config.d/netconf = 3987

Regards!

...JRF...

[James R. Ferguson]

Hi Scott:

If you want to limit the output only to regular files, amend my original post to:

# perl -le 'die unless @ARGV;for (@ARGV) {print join " = ", $_,(stat(_))[7] if (-f $_)}' /etc/hosts/ /etc/services /etc/rc.config.d

...returns (only):

/etc/hosts/ = 606
/etc/services = 9946

...since '/etc/rc.config.d' is a *directory*. My first variation treats files and directories equally and reports their byte-sizes.

Regards!

...JRF...

[A. Clay Stephenson]

... and if you already have the file open, the fstat() system call is a little easier because it uses the file descriptor rather than the pathname as the 1st argument.

[Ralph Grothe]

If you can live with perl
there's even an easier way than to stat().
I wonder why people always forget about
the simple -s test.
E.g. this should print a size sorted list of all files in /tmp
(n.b. usually one would further refine by excluding the single dot references as well as directories etc.)

$ perl -e 'opendir TMP,"/tmp";print map{sprintf"%30s:%10u\n",$_->[0],$_->[1]}sort{$a->[1]<=>$b->[1]}map[$_,-s "/tmp/$_"],readdir TMP;closedir TMP'

[inventsekar_1]

Is there an easy way to programmatically read file size either from command line or via script:

"wc -c filename" will "Report the number of bytes in each input file".

is this fine?
[ofcourse, u need to cut the result of wc]

[Arturo Galbiati]

wc -c
HTH,
Art

[Steven E. Protter]

Shalom,

Change:
wc -c filename

filename=
SIZE=$(wc -l $filename)
echo $SIZE

SEP

[inventsekar_1]

SEP,

filename=
SIZE=$(wc -l $filename)
echo $SIZE

should be
filename=
SIZE=$(wc -c $filename)
echo $SIZE

isnt?

[-l gives the number of lines
-c gives the number of bytes]