Hi, I am having a problem determinig how to write an awk script that will do the following. I have a file and in that file I am looking at a certain field. This field will contain one of tow things. All numbers or a cobination of numbers and chars. If I see chars in the field I want to append 10 space to the left of the field. If I see that the field has all numbers, then I ignore it and go to the next record. I can hanlde the apeend of 10 spaces etc. What I do not know is if there is a way to tell a number from a char in a text file. Is there a way to know that "2" is a number and "B" is a letter?

One of the possible solution. #!/bin/ksh Let_Field=123456 if [ "$Let_Field" -eq "$Let_Field" 2>/dev/null ] then echo "It Is Numeric" echo "Do Whatever you want to do" else echo "It Is Not Numeric" echo "Do Whatever you want to do" fi

Do an awk search for [a-zA-Z] (i.e) /[a-zA-Z]/

How does the awk search help? If I am looking at a string of 7 characters in field 3 or $3. And I search for [a-zA-z] I can return $3 with 10 spaces. But I only want 10 spaces at the back end of the 7 char string if chars are found. So How do I do the if in awk? I.E. if the filed contains "12T4567" then I want to return to a file "12T4567 " but if the field is "1234567" I want to return "1234567". SO I just do not want to find only the rcords with chars I want all records. Just want to manipulate the ones with chars.

Sorry the above example i response #3 should be I.E. if the filed contains "12T4567" then I want to return to a file "12T4567 " but if the field is "1234567" I want to return "1234567". SO I just do not want to find only the rcords with chars I want all records. Just want to manipulate the ones with chars.

Hi Tom, I hope I understood it correctly. try: awk '{ if ( $2 ~ "[a-z]" ) print $1,$2"_six_balnk",$3 else print $1,$2,$3 }' file file containes e.g.: abcs adrfe dbea abcs a dbea abcd 123 dbea abdc 1d2 dbea output would be: abcs adrfe_six_blank dbea abcs a_six_blank dbea abcd 123 dbea abdc 1d2_six_blank dbea Hope it helps NO RISK no fun Frank

Why get awk involved in this problem? Simple pattern match can be done right in the script. Sample: #!/bin/ksh case $1 in +([0-9]) ) print "It's numeric" ;; * ) print "It ain't numeric" ;; esac Modify the above logic to evaluate a field from an input file and you have it. Jerry