>Hello People,
> I have a table set up with the following fields: date(D),fund_id(C,10),shareclass(C,10),performance(N,10,3). For every unique fund_id there are anywhere from 1 to 25 share classes with performance history over a range of dates. I want to set up a way to update the performance values for a particular fund in a tabular format. For example, if FundA has 4 share classes and is chosen it would look like:
>
>Date Class1 Class2 Class3 Class4
>01/01/01 2.2 2.0 2.3 1.9
>01/02/01 -1.5 -1.3 -1.2 -1.7
>.
>.
>.
>05/25/01 5.3 4.9 5.0 5.1
>
>Is there a way to do this with a view created from a self-join? Should I set the table up differently. There are thousands of funds, so the file would get awfully big if I had a field for each possible share class. I am new to SQL so I apologize for my ignorance if the answer to my question is obvious or if I haven't explained my problem succinctly.
>
>Thanks in advance!
>
>- Tony

I would think you'd want to split the table into two: a fund table and a class table, like this:

create table FUND (date(D),fund_id(C,10))

create table FUNDSHARE (fund_id(c,10), shareclass(C,10),performance(N,10,3))

(no guarantees the syntax is correct)

The FUNDSHARE will be a child of the FUND table, linked via the common FUND_ID field. Yes, the FUNDSHARE table will get big but this is the best way to do this, since you don't know how many shares could belong to a fund. Yes, you could "best guess" the maximumn and set up that many columns in the fund table, but as soon as the share count exceeds your best guess - and it will - you have a potentially serious maintenance issue.