Re: On-Line Schedule of Classes -Reply

sfahey@niu.edu
Thu, 22 Jan 98 16:47:32 -0600

Nice job! Wonderful description.
I'm not real familiar with Access so bear with me.
Why couldn't you create the tailored database using SQL
Server and let SQL Server manage it? Does Access give
you something special that SQL Server does not?

Sean

In <Pine.SOL.3.91.980122154739.6998B-100000@oats>, on 01/22/98
at 04:38 PM, David Benham <z935418@oats.farm.niu.edu> said:

>We use all three tools actually. The data on the SQL server is not
>organized the way it would need to be to allow some detailed and complex
>queries. So we built a fairly customized database in Access from the SQL
>server. But we couldn't do large searches on the Access DB because Access
>DBs are VERY SLOW, so we still needed to use SQL server to help speed
>things up. The average query time to the SQL server(over the network) is
>less than 5 ms. The average search on a local Access DB took well over
>100 ms. Big difference, especially when you are doing hundreds if not
>thousands of queries for every search.

>So its a trade off. The slower MS Access DB we use could do more detailed
> searches because we could tailor the structure of the DB to our needs.
>At the same time, the Access DB was slow. The SQL server is much faster,
>but the data wasn't formatted the way we needed it. So we needed to
>combine both. When we needed complex searches done, we used information
>from an Access DB. When speed was needed we went to the SQL server.

>Cold Fusion, besides from allowing us to connect to the datasources via
>ODBC, provided some control logic that allowed us to break the search up
>across two different databases and CF also helped in providing the
>formatted output.

>Giving Cold Fusion access to the two different databases through ODBC
>wasn't too difficult. There were several problems getting everything to
>work on NT after it was developed and tested on a win95 machine, but they
>were dealt with.

>Like you said, the data wasn't stored the way we needed it, but that is
>due mostly to the fact the data wasn't organized for us, it was organized
>to be used in a different manner. But with some clever logic(?) :) and
>the Cold Fusion program, we were able to get it to work.

>-dave

>**************************************************************************
>* David Benham
>* * dbenham@niu.edu
>* * http://mp.cs.niu.edu/~z935418
>*
>**************************************************************************