for each file if file name like xx*
for each line in a file
if substring(3,6) found in another txt file[ the file just have 1 column but have multiple line]
output to file-a( filename = orginal file + _a)
else
output to file-b( filename = orginal file + _a)
end
Next Line

Assuming you can use bash
See here for how to loop
See here for substring manipulations
use ">>" or ">" to output to file

I am new for sh script , i have write some code but cann't compile , please help for seeing what problem? It seem to cannot read another file in awk and I don't confirm....

"XXXX.wip.sh" 24 lines, 471 characters
>findcust.wip.sh
Syntax Error The source line is 7.
The error context is
while >>> read <<< line
awk: 0602-502 The statement cannot be correctly parsed. The source line is 7.
Syntax Error The source line is 8.


awk -v '
BEGIN {FOUND = "N";}

{if (match($0, /\!\$....../))
{checkno = substr($0,3,6);
FOUNDINCUST = "N" ;
while read line
do
{if (FOUNDINCUST="N") && (checkno == $line)
FOUND = "Y";
FOUNDINCUST ="Y";
print $0 ;
}
done < "custfilelist.dat" ;
}
}
{if (match($0, "%%RECORDEND"))
FOUND = "N";}
{if (FOUND =="Y")
print $0 >> "atest.dat" ;}
{if (FOUND =="N")
print $0 >> "btest.dat" ;}
' $SOURCEFILE/TEST*.dat

You're mixing awk with sh.

thanks for your reply , it means that I cann't have sh script inside the awk call ? then what should I do to do the objective , should i just use the sh script , not use the awk to do it , any code sample that is doing similar things ??