I am working with statistic analysis in SPSS but i find a relatinal database a good place to store information. VFP is a good choice since I have been using it a lot for other purposes. I use to get data in ASCII-files and read them in to SPSS for further analysis. Hereby I have to treat blanks as system missing wich I tell SPSS when I read the file. If I don't, I will end up calculating means on cases where data is missing (no good). However, when I read the ASCII-file in to VFP (5.0) blanks look like blanks when I browse, but are treated as zero. I have managed to edit thoses values to .NULL. value by value, but since I have thosands of cases and hundreds of variables (records and fields) this is a tedeous job (impossible). Since VFP treats my blanks as zero I am not able to replace the blanks with .NULL. with the replace command since there is a difference between my blanks and my zeroes and VFP treats them all as zeroes. Another funny thing is that when I edit my records manually to contain .NULL. by using CTRL+0. VFP just ignore those .NULL.-vallues. That means they are treated the correct way when I calculate means with the AVG() funktion, but when I want to calculate frequencies with COUNT() they are allways 0 even if I know they are hundreds. I have to be able to tell how many people did not respond.
I would be gratefull for help on this topic since tecnical support don't know.
Per Strid
Nothing is impossible, it's just more or less complicated!