Hi All. I'm developing a payroll system that imports data from large governmental agencies, extracts the useful info and works with that.

The Raw Import data is quite large (about 20MB/month), whereas the resulting normalized data is much smaller (under 1MB/month).

For auditing reasons, I'd like to keep all the data online. However, I'm trying to decide whether to keep the Raw Import data table in the main file, or in a separate file. I like the simplicity of the one-file design in terms of backups, managing user accounts ,etc. However, I'm concerned with performance and stability issues. I'd like this solution to handle at least 30 years of data (which means the single-file solution would grow to well over 5GB).

I know that FM7 claims to handle multi-gigabyte files with ease -- but I'm still a little nervous. Of course, in 10 years, we'll probably have multi-terabyte hard drives, so maybe 5GB of data is not that much...

Any advice?

I really don't think it's going to matter all that much. And if you change your mind, it probably won't be that hard to move the table into or out of the main file, since I assume you don't have much in the way of interface, scripts and relationships to worry about in a raw data file.