Hello All, The script need to read a file, which contains such as follows: Mary age22 address1 Peter age33 address2 eof then all data put into variable Output like : name = Mary age = age22 address = address1 name = Peter age = age33 address = address2

awk '$2 = substr($2,4) { printf "name = %s\nage = %s\naddress = %s\n\n", $1, $2, $3 }' infile

If you want output formatting, David gave you a solution for that.  If you need to load the fields into variables so that you can work with the variables, the script below provides an example: while read name age addr other do # echo name=$name age=$age addr=$addr printf "name=%-10s age=%-6s addr=%s\n" $name $age $addr done < infile

Thanks. I have another question. If I want to get exactly field size of name, age and address. What should I do? for example: input name = char 5 age = char 2 addr = char 40 while true $read name | cut -d -f,5 $read age | cut -d -f,2 $read addr | cut -d -f ,40 do # echo name=$name age=$age addr=$addr printf "name=%-10s age=%-6s addr=%s\n" $name $age $addr done < infile Can I do like that? Thanks and Regards, Apple

If you are asking how to pull variables from each line of your file based upon fixed column positions, then I believe you have the general idea of it. In the ksh script below, the while-loop reads each line of infile into the $line variable, then I show two different ways to split up $line based upon fixed column positions. IFS="" while read line do name=$(expr "$line" : "\(.....\)") age=$(expr  "$line" : ".....\(..\)") addr=$(expr "$line" : ".......\(.*\)") #name=$(echo $line | cut -c1-5) #age=$(echo $line | cut -c6-7) #addr=$(echo $line | cut -c8-) printf "name=%-6s age=%-3s addr=%s\n" $name $age $addr done < infile