-----BEGIN PGP SIGNED MESSAGE-----
Sunday, January 31, 1999
This is File::Sort 0.18, for sorting files similarly to sort(1). Written
primarily for MacPerl users who do not have sort(1) and because of memory
limitations cannot sort files in memory, but works on all perls.
See HISTORY below for changes.
This archive can always be obtained from:
http://perl.com/CPAN/authors/id/CNANDOR/
http://perl.com/CPAN/modules/by-module/File/
Please let me know how well it does(n't) work, and any changes you'd
like to see.
#=======================================================================
NAME
File::Sort - Sort a file or merge sort multiple files
SYNOPSIS
use File::Sort qw(sort_file);
sort_file({
I=>[qw(file1_new file2_new)],
O=>'filex_new',
V=>1, Y=>1000, TF=>50, M=>1, U=>1, R=>1, N=>1,
});
sort_file('file1','file1_new',1,1000);
DESCRIPTION
WARNING: This is probably going to be MUCH SLOWER than using sort(1)
that comes with most Unix boxes. This was developed primarily because
some perls (specifically, MacPerl) do not have access to potentially
infinite amounts of memory (thus they cannot necessarily slurp in a
text file of several megabytes), nor does everyone have access to
sort(1).
Here are some benchmarks that might be of interest (PowerBook G3/292
with 160MB RAM, VM on, and 100MB allocated to the MacPerl app). The
file was a mail file around 6MB. Note that once was with a CHUNK value
of 200,000 lines; Unix systems can get away with something like that
because of VM, while Mac OS systems cannot, unless you bump up the
memory allocation as done below. So inevitably you will get much
better performance with large files on Unix than you will on Mac OS.
C'est la vie.
Note that tests 2 and 3 cannot be performed on the given dataset when
MacPerl has a small amount of memory allocated (like 8MB). But when
MacPerl has 8MB allocated, the results for tests 1 and 4 are about the
same as when MacPerl has 100MB allocated, showing that the module is
doing its job properly. :)
NOTE: `sort` calls the MPW sort tool here, which has a slightly
different default sort order than `sort_file' does.
#!perl -w
use File::Sort qw(sort_file);
use Benchmark;
timethese(10,{
1=>q+`sort -o $ARGV[0].1 $ARGV[0]`+,
2=>q+open(F,$ARGV[0]);open(F1,">$ARGV[0].4");@f=<F>;print F1 sort @f+,
3=>q+sort_file({I=>$ARGV[0],O=>"$ARGV[0].2",Y=>200000})+,
4=>q+sort_file({I=>$ARGV[0],O=>"$ARGV[0].3"})+,
})
Benchmark: timing 10 iterations of 1, 2, 3, 4...
1: 185 secs (185.65 usr 0.00 sys = 185.65 cpu)
2: 152 secs (152.43 usr 0.00 sys = 152.43 cpu)
3: 195 secs (195.77 usr 0.00 sys = 195.77 cpu)
4: 274 secs (274.58 usr 0.00 sys = 274.58 cpu)
That all having been noted, there are plans to have this module use
sort(1) if it is available. Still.
WARNING Part Deux: This module is subject to change in every way,
including in the fact that it exists. But it seems much less subject
to change now than it did at first.
There are two primary syntaxes:
sort_file(FILEIN, FILEOUT [, VERBOSE, CHUNK]);
This will sort FILEIN to FILEOUT. The FILEOUT can be the same as the
FILEIN, but it is required. VERBOSE is off by default. CHUNK is how
many lines to deal with at a time (as opposed to how much memory to
deal with at a time, like sort(1); this might change). The default for
Y is 20,000; increase for better performance, decrease if you run out
of memory.
sort_file({
I=>FILEIN, O=>FILEOUT, V=>VERBOSE,
Y=>CHUNK, TF=>FILE_LIMIT,
M=>MERGE_ONLY, U=>UNIQUE_ONLY,
R=>REVERSE, N=>NUMERIC,
D=>DELIMITER, F=>FIELD,
S=>SORT_THING,
});
This time, FILEIN can be a filename or an reference to an array of
filenames. If MERGE_ONLY is true, then `File::Sort' will assume the
files on input are already sorted. UNIQUE_ONLY, if true, only outputs
unique lines, removing all others.
FILE_LIMIT is the system's limit to how many files can be opened at
once. A default value of 40 is given in the module. The standard port
of perl5.004_02 for Win32 has a limit of 50 open files, so 40 is safe.
To improve performance increase the number, and if you are getting
failures, try decreasing it. If you get a warning in `_writeTemp',
from the call to `_getTemp', you've probably hit your limit.
If given a DELIMITER (which will be passed through `quotemeta'), then
each line will be sorted on the nth FIELD (default FIELD is 0). If
sorting by field, it is best if the last field in the line, if used
for sorting, has DELIMITER at the end of the field (i.e., the field
ends in DELIMITER, not newline).
SORT_THING is so you can pass in any arbitrary sort thing you want,
where $SORT is the token representing your $a and $b. For instance,
these are equivalent:
# {$a cmp $b}
sort_file({I=>'b', O=>'b.out'});
sort_file({I=>'b', O=>'b.out', S=>'$SORT'});
# {(split(/\|/, $a))[1] cmp (split(/\|/, $b))[1]}
sort_file({I=>'b', O=>'b.out', D=>'|', IDX=>1});
sort_file({I=>'b', O=>'b.out', S=>'(split(/\\|/, $SORT))[1]'});
SORT_THING will still need R and N for reverse and numeric sorts.
Note that if FILEIN does not have a linebreak terminating the last
line, a native newline character will be added to it.
EXPORT
Exports `sort_file' on request. `sortFile' is no longer the function
name.
BUGS
None! :) I plan on making CHUNK and FILE_LIMIT more intelligent
somehow. I did make the default for CHUNK larger, though.
Also, I will have the module use sort(1) if it is available.
THANKS
Mike Blazer <blazer@mail.nevalink.ru>, Vicki Brown <vlb@cfcl.com>,
Gene Hsu <gene@moreinfo.com>, Andrew M. Langmead <aml@world.std.com>,
Brian L. Matthews <blm@halcyon.com>, Rich Morin <rdm@cfcl.com>,
Matthias Neeracher <neeri@iis.ee.ethz.ch>, Miko O'Sullivan
<miko@idocs.com>, Tom Phoneix <rootbeer@teleport.com>, Gurusamy
Sarathy <gsar@activestate.com>.
AUTHOR
Chris Nandor <pudge@pobox.com> http://pudge.net/
Copyright (c) 1998 Chris Nandor. All rights reserved. This program is
free software; you can redistribute it and/or modify it under the same
terms as Perl itself.
HISTORY
v0.18 (31 January 1998)
Tests 3 and 4 failed because we hit the open file limit in the
standard Windows port of perl5.004_02 (50). Adjusted the default
for total number of temp files from 50 to 40 (leave room for other
open files), changed docs. (Mike Blazer, Gurusamy Sarathy)
v0.17 (30 December 1998)
Fixed bug in `_mergeFiles' that tried to `open' a passed
`IO::File' object.
Fixed up docs and did some more tests and benchmarks.
v0.16 (24 December 1998)
One year between releases was too long. I made changes Miko
O'Sullivan wanted, and I didn't even know I had made them.
Also now use `IO::File' to create temp files, so the TMPDIR option
is no longer supported. Hopefully made the whole thing more robust
and faster, while supporting more options for sorting, including
delimited sorts, and arbitrary sorts.
Made CHUNK default a lot larger, which improves performance. On
low-memory systems, or where (e.g.) the MacPerl binary is not
allocated much RAM, it might need to be lowered.
v0.11 (04 January 1998)
More cleanup; fixed special case of no linebreak on last line;
wrote test suite; fixed warning for redefined subs (sort1 and
sort2).
v0.10 (03 January 1998)
Some cleanup; made it not subject to system file limitations;
separated many parts out into separate functions.
v0.03 (23 December 1997)
Added reverse and numeric sorting options.
v0.02 (19 December 1997)
Added unique and merge-only options.
v0.01 (18 December 1997)
First release.
VERSION
Version 0.18 (31 January 1998)
SEE ALSO
sort(1).
#=======================================================================
- --
Chris Nandor mailto:pudge@pobox.com http://pudge.net/
%PGPKey = ('B76E72AD', [1024, '0824090B CE73CA10 1FF77F13 8180B6B6'])
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0 for non-commercial use <http://www.pgp.com>
Charset: noconv
iQCVAwUBNrSCtChcZja3bnKtAQHF3wP9FFvcR4VCPaE8zqwMURnOsIE1O4Ci4Hf7
yoeElcIcFn11G0JfCsXNzJBrIfV6e5X2mB7YLlGpOi63/0oMfvq8k0qLpf6DLv+N
oGP/ud4DX4uZ58gl9zpJEkQPEK9b2vW+hZuo2TqZS3ZwIUK5Z9wGErrjRj+vAveW
EYCpdG8aJxU=
=eQ6Z
-----END PGP SIGNATURE-----