Shell Programming and Scripting

Hi,
I was wandering would it be possible to split the record dynamically based on the certain values, for an instance

i have a file with record with predefined split value i.e 10

col1 col2 col3 col4
------------------------
aaaa bbbb 2 44aaaabbbb55cccddd1110
mmn xnmn 3 808klmbbhm90bvnbvnbvnvn


so for the first row col3 has 2, so col4 need to split of 2 parts of 10 chars each,likewise for second record col4 need to be split into 3 parts of 10 chars

so the final normalised output should look like

col1 col4
----------------
aaaa 44aaaabbbb
aaaa 55cccddd11
mmn 808klmbbhm
mmn 90bvnbvnbv

would this be possible in shell scripting ?

Thanks in advance

Look at your example. Why didn't your 2nd input record generate a 3rd output record? And I assume that when there aren't enough input characters, you just use blanks?

Hi,
Sorry I missed out while typing, for the second input there should have been one more row,

col1 col4
----------------
aaaa 44aaaabbbb
aaaa 55cccddd11
mmn 808klmbbhm
mmn 90bvnbvnbv
mmn nbvnv

This script:

Code:

#! /usr/bin/ksh

exec < data1
typeset -L10 leftpart

while read col1 col2 col3 col4 ; do
        n=col3
        while ((n)) ; do
                leftpart=$col4
                typeset -R$((${#col4} - 10)) rightpart=$col4
                col4=$rightpart
                ((${#col4}<10)) && col4="$col4            "
                echo $col1 $leftpart
                ((n=n-1))
        done
done
exit 0

results in:

Code:

aaaa 44aaaabbbb
aaaa 55cccddd11
mmn 808klmbbhm
mmn 90bvnbvnbv
mmn nvn

Awk:

Code:

{ for (i=0; i<$3; i++)
    print $1, substr($4,i*10+1,10)
}

By dynamically splitting did you mean the value '10' can also change? If yes, you can use the awk statement provided by futurelet in a shell script, like this:

split_file.sh:

awk -v a=$1 '{
for (i=0;i<$3;i++)
print $1,substr($4,i*a+1,a)}' file_to_be_split.txt

This script takes the no. of characters in the substring as an argument. Yo can call it like this :
sh split_file.sh 10

Thanks a ton guys, no. of chars is fixed, so mere awk will do. Indeed quite a powerful one.

Thanks again