ファイルを行ごとにランダムにソートして、指定した数のファイルに分割する。randomly sort and split given file in given number.

randomsplit.sh

#!/bin/bash

# usage:
# ./randomsplit.sh [FILE] [division number]
# 

# Check the number of parameters
if [ $# -ne 2 ]; then
  echo "usage: ./randomsplit.sh [FILE] [division number]" 1>&2
  exit 1
fi

file=`echo $1` # filename
fold=`echo $2` # division number

# Randomsort for given file
cat $file | awk 'BEGIN { srand() } { print rand() "\t" $0 }' | sort -n | cut -f 2- > tempfile

# Count lines of given file
line_count=`wc -l $file|sed 's/^ *//g'|cut -f 1 -d ' '`

# Split randomized file
split -l `expr $line_count / $fold` tempfile splited

# Add remained lines to each file
remainder=`expr $line_count % $fold`
if `test $remainder -ne 0`; then
  for i in `seq 1 $remainder`
  do
    writefile=`ls | grep 'splited' | head -n $i| tail -n 1`
    tail -n $i tempfile | head -n 1 >> $writefile
  done
fi

# Delete unnecessary files
filecount=0
for i in `ls | grep 'splited'`
do
  if `test $filecount -lt $fold`; then
    filecount=`expr $filecount + 1`
  else
    rm $i
  fi
done
rm tempfile