Combining awk and csum to hash a field - unix

Question

I have pipe-delimited text files that requires an MD5 hash of a particular field, or set of fields. Because I'm on AIX and have to use the csum function, I don't think I can simply pass the file and a hashing function to awk to do it in one fell swoop.
So I'm writing a script that reads through each line, passes the to-be-hashed field to csum, then drops the result back in as a replacement via a gsub. 99% of the time it seems to work OK, but sometimes something goes afoul because the gsub replaces something it shouldn't.
#!/bin/ksh
rm $2 #Get rid of output file
while read line; do #loop through each line
MYFIELD=$(echo "$line" | cut -d "|" -f 6); #push the 6th field into a var
MYHASH=$(echo $MYFIELD | csum -h MD5 -); #csum will hash a string only on the stdin
echo $line | sed -e "s/$MYFIELD/${MYHASH}/g" >> $2 #gsub replaces, but not always what we want
done < $1 #read in the input file
I think instead I could use awk to update the field. But it's beyond me how to do that one line at a time. Ideally I would like to have a script that would allow me to pass two mandatory parameters (infile and outfile) and then any number of field positions that would get hashed and replaced. A la
foo infile.txt outfile.txt 2 6 12
Which would read in infile.txt, hash fields 2, 6, and 12, and write out to outfile.txt.
Your suggestions would be most appreciated

Answer 1

What about doing it with awk?
Instead of
echo $line | sed -e "s/$MYFIELD/${MYHASH}/g" >> $2 #gsub replaces, but not always what we want
You can use
old=$MYFIELD; new=$MYHASH; echo $line | awk -F"|" -v o="$old" -v n="$new" '{OFS=FS} sub(o, n, $6) {print}' >> $2
Basically what we do is:
old=$MYFIELD; new=$MYHASH We assign the parameters to be sent to awk.
echo $line We output the line so that awk can get it.
In awk,
-F"|" define | as field separator.
-v o="$old" and -v n="$new" let awk work with variables $old and $new naming them o and n respectively.
{OFS=FS} - define the delimiter between fields. It could also be OFS="|", but this way we indicate awk to use the same we defined on -F="|". It is more flexible to keep the field separator in case it changes.
sub(o, n, $6) replaces the text on variable o (that is, $MYFIELD) with text on variable v (that is, $MYHASH), but just on field 6.
print the whole line with substituted text
This worked for me in the example you gave on comments:
old="hashit"; new="WE_DID"; echo "donthashit|foo1|bar1|foo2|bar2|hashit" | awk -F"|" -v o="$old" -v n="$new" '{OFS=FS} sub(o,n,$6) {print}'
donthashit|foo1|bar1|foo2|bar2|WE_DID
Hope it helps.
Edit
I found a way to pass variables to awk easily: -v o=${variable_name}
This way, the solution can be:
echo $line | awk -F"|" -v o=${MYFIELD} -v n=${MYHASH} '{OFS=FS} sub(o, n, $6) {print}' >> $2