On Jan 14, 2008 11:28 PM, mark pirogovsky <[EMAIL PROTECTED]> wrote: > It is really hard to give you some idea without knowing more of you > problem... but your proposed implementation does carry a lot of > overhead - for every number you have two extra fields, not to mention > the database internal structures. >
Basically, I am trying to connect SQLite up to Pure Data (a dataflow programming language), and I am creating arbitrary lists of floating point numbers that I would like to be able to store as sets. The SQL code from my original message is what I am currently using for the array stuff. Just one big table with each element tagged for grouping and another for their order. It sounds like what you described would be great, if it were only in SQLite. I had thought about the idea of creating everything as a long string, separated by a space, and storing them as a group. I just wasn't sure if this was the "accepted" way to do it in SQL. Thanks for your suggestions. Mike
