manipulating time series data

        I am new to ORACLE and would appreciate advice on the best way to do the following.

        I am trying to set up a table that contains a very large number of trials where each trial produce a brainwave floating point value for a large number (but probably not over 2000) of points over time. The brainwave values need to be manipulated individually.

        As I see it, I have the following options.

1. I tried setting up a table with a record for each time point but got almost 2 million records for just one file. I will be dealing with hundreds of files. I am concerned that this option would not be feasible because of the amount of disk space involved.
2. I could store the values of each trial in a column of type RAW. But, as I understand it, I would not be able to use SQL queries on the values individually, i.e. select trial_id, time_pt from tableX where value > 10.145. Is this correct? Would I need to write C routines to do this? Would this turn the whole process of querying into a very slow one?
3. I could have a pointer to a file containing the data points. But the retrieval time would be horrendous. I see this as the least feasible option.

        What do you all think? Which would be the best route to take? Thanks very much for taking the time to consider my problem.

Jenny Barrett Received on Mon May 16 1994 - 18:36:09 CEST