i'm trying to make a script that split a file divided by nbytes. I already note this but i want to use head and tail and not using split as me.
#!/bin/sh
if [ $# -eq 0 ];then
exit 1
fi
if [ $# -eq 1 ];then
exit 1
fi
if [ $2 -eq 0 ];then
exit 1
fi
if [ ! -f "$1" ];then
exit 1
fi
split -d -b 1024 bigfile.bin bigfile.bin.
It giving this :
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.00
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.01
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.02
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.03
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.04
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.05
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.06
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.07
-rw-rw-r-- 1 madushan madushan 1024 déc. 10 17:34 bigfile.bin.08
-rw-rw-r-- 1 madushan madushan 784 déc. 10 17:34 bigfile.bin.09
You can implement a light-weight version of split with bash (+head, +tail). However, it will not be very efficient, as you will need to read the file times, where N=totalsize/nbytes. For small files, the overhead is small, for large file, very expensive.
nbytes=1024
file=bigfile.bin
k=0
i=0
while tail --bytes=+$((nbytes*i)) < $file | head --bytes=$nbytes > $file.work ; do
# Stop unless segment has data
[ -s "$file.work" ] || break
let i++
echo "Segment: $i"
mv "$file.work" "$file.$i"
done
rm -f $file.work
If OK with just using head, possible to make is more efficient for large files. It will read the input only once, with no need to re-read anything.
nbytes=1024
file=bigfile.bin
k=0
i=0
(
while head --bytes=$nbytes > $file.work ; do
[ -s "$file.work" ] || break
let ++i
mv "$file.work" "$file.$i"
done
) < $file
rm -f $file.work
Also consider using 'dd', which has more powerful logic for large files.