Hi,
The way you're trying to assign values to the custid variable will not work because the AWK output will be regarded as one string, as you noticed. I think you should choose the environment in which to do both the assingment and further processing of the array. And if it's impossible to do the complete handling of the array in AWK you'd better try to fill the array in shell script.
Since your AWK program only reads the first field of every record you could do the same with 'sed' for instance, like:
Code:
#!/usr/bin/ksh
i=0
for field in `cat custid_file_name | sed 's/,.*$//'`
do
var[$i]=`echo $field`
((i=$i+1))
done
j=0
while [ $j -lt $i ]
do
print ${var[$j]}
((j=$j+1))
done
With this you've loaded the array completely in the shell environment. And for loading the same array in AWK (which is not available in the shell environment) I would do it like:
Code:
#!/usr/bin/ksh
awk -F "," '{ var[NR] = $1 }
END { for(i = 1; i <= NR; i++)
print var[i]
}' custid_file_name
where NR is the builtin variable for recordnumber read and the -F flag sets the fieldseparator.
Regards
dBforums
>
How to create Array variable?
01-26-07, 12:15
Linear Mode
