CODEWITHSUNDEEP
linuxbashunix
alpha_cod
alpha_cod

Reputation: 2033

Split file into unequal chunks in Linux

I wish to split a large file (with ~ 17M Lines of Strings) into multiple files with varying number of lines in each chunk. Would it be possible to send in an array to the 'split -l' command like this:

[
 1=>1000000,
 2=>1000537,
 ...
]

so as to send those many number of lines to each chunk

Upvotes: 8

Views: 3385

Answers (3)

that other guy
that other guy

Reputation: 123690

Use a compound command:

{
  head -n 10000 > output1
  head -n   200 > output2
  head -n  1234 > output3
  cat > remainder
} < yourbigfile

This also works with loops:

{
  i=1
  for n in 10000 200 1234
  do
      head -n $n > output$i
      let i++
  done
  cat > remainder
} < yourbigfile

This does not work on OS X, where head reads and discards additional output.

Upvotes: 12

sotapme
sotapme

Reputation: 4903

You could use sed by getting another script to generate the sed commands for you.

# split_gen.py
use strict;
my @limits = ( 100, 250, 340,999);
my $filename = "joker";

my $start = 1;
foreach my $end (@limits) {
    print qq{sed -n '$start,${end}p;${end}q' $filename > $filename.$start-$end\n};
    $start = $end + 1;
}

Run thus perl split_gen.py giving:

sed -n '1,100p;100q' joker > joker.1-100
sed -n '101,250p;250q' joker > joker.101-250
sed -n '251,340p;340q' joker > joker.251-340
sed -n '341,999p;999q' joker > joker.341-999

If you're happy with the command then you can 

perl split_gen.py | sh

Then enjoy the wait as it may be slow with big files.

Upvotes: 1

Jim Lewis
Jim Lewis

Reputation: 45135

The split command does not have that capability, so you'll have to use a different tool, or write one of your own.

Upvotes: 1

Related Questions