I need to split a long varible which is a whole line read from a file into fields and store them in an array, the fields are delimited by pipe and a field may contain white spaces.

I tried the following concept test and it has problem with field 5 which contain a space, appearently so because "set -A" treats the space as a field delimiter.

Is there any other better way that the script can receive awk output and store them in an array with white spaces perserved?

TIA.


#!/bin/ksh

var="word1#word2|word3/word4|word5.word6|word7_word8|word9 word10|word11|word12"

set -A varray `echo "$var"| awk '{z=split($0,flds,"|")
for(i=1;i<=z;i++)
print flds[i]}'`

echo ${varray[0]}
echo ${varray[1]}
echo ${varray[2]}
echo ${varray[3]}
echo ${varray[4]}
echo ${varray[5]}
echo ${varray[6]}

---------
script output:

$test_read.ksh
word1#word2
word3/word4
word5.word6
word7_word8
word9
word10
word11

White spaces are the default field separator for awk

/home/jmcnama>  var="word1#word2|word3/word4|word5.word6|word7_word8|word9 word10|word11|word12"
/home/jmcnama> set -A arr `echo "$var"| awk -F'|' '{for(i=1; i<=NF; i++){printf("%s ", $i)}}'`
/home/jmcnama> echo ${arr[*]}
word1#word2 word3/word4 word5.word6 word7_word8 word9 word10 word11 word12
csadev:/home/jmcnama>

or a bit simpler:

#!/bin/ksh

var='word1#word2|word3/word4|word5.word6|word7_word8|word9 word10|word11|word12'

IFS='|'; set -A arr $(echo "$var")
i=0
while (( i <= 6 ))
do
    echo "i=>[$i] (${arr[$i]})"
    ((i=i+1))
done

yes, these worked. Thanks a lot.