I have stumped 2 unix scripters already

I have hundreds even thousands of files where each file has a ^ field separator and a ~ line separator. there are no carriage returns or line feeds at all in these files. So the file is basically just one big long line of text. I need to count the number of times the text "PO1" appears in each of these files in this directory. I am running the script or I can run from the command line, in the directory where these files reside, so, I have tried the awk, grep, sed commands, nothing works, I have received assistance from 2 very experienced unix script developers and they are stumped also. any ideas on how to get this done?

My guess is that you have not stumped anyone here - and if you have they are not very good at what they do...
That aside, what have you tried so far that has not worked? What has it produced?

sed 's/~/\n~/g' < filename > outputfile

i tried this to get a new line inserted into the file. i got 0 output, if i can get the tilde to become a new line then awk or grep will work.

This “works” if you can get the PO1’s on their own lines. You could do that with a sed statement.

find . -type f –name -exec grep -i -c "PO1" {} \; > a1
awk '{total += $1}; {print total}' a1 > a2
tot=`tail -1 a2`
echo $tot
rm a1 a2

A file formatted as you specified (albiet a bit smaller)

Counting teh number of "PO1" occurances in it
Caveat: If there are two PO1 items within a single field, they will not be counted twice. This essentially counts the number of fields in the file that contain the text "PO1". I figured that was ok considering your grep example.

for afile in `find . -type f -print`

do

tr '~' '\n' < $afile | tr '^' '\n' | egrep "PO1" | wc -l > countlines.txt

done