Thanks to all answers on my question! I've finished my work.

In my actual application, I have two files: one (file A) is the original data file, the other one(file B) is some kind of analysis result for the original data file. For some reason, the record order in file B is not same with file A.

Now, I need to do some analysis based on each record in file A and B. So, I decided to create an index for file B, access file A sequencially, and seek current record's content on file B.