Note: This may have something to do with the fact that half of my posts might be invalidated if we insisted that I actually provide technical insight on everything I say. :)
In any event, if you treat a flat file as a database table, each line of input becomes a database "row." In that context, you can identify "fields" in that table by their relative position in the line (fixed width), or by split /x/ on some character (delimited). If you sysread() your data, I guess you could treat it as binary. Does anyone do this any more? I don't think I've done it since I worked (in C) under VAX/VMS, and it's mostly been supplanted in the database world by BLOBs (binary large objects).
Having now said this, if you are trying to get a database job that isn't entry level, you're probably overmatched (as Fletch has stated). But that said, I never object to people wanting to grow in knowledge. If you show a willingness to make the effort to learn, it may just get you where you want to be.