Re: What type of DB model is this?
Date: Thu, 16 Aug 2001 10:46:15 GMT
Message-ID: <XvNe7.433$ms2.43794_at_news1.cableinet.net>
Wolfgang,
Thanks for sticking with it.
>>Next question is the programming language you want to use <<
Java.
I was pretty much intending on developing my own data structure (as opposed to using a database product) but my problem is how to go about this in the most efficient way.
Last time I did something with a similar data structure to this, I created 1 flat file which contained:
- a field of 200 chars for each of the top level items
- 12 fields of 140 chars for the each of the second level 2nd level items
- 10 fields of 80 chars for each the 3rd level items.
and used simple arithmetic to work out where the start of each field was located within the file.
The problem was that the second level may have contained only 1 field of data (12 was the maximum that could be stored) and it's third level may also have contained only 1 field. But my database would have stored 12 * 10 = 120 fields at the third level of which only 1 was used. This left me with a database which was 32MB in size, and probably only about 5% of it contained data. But I had to allow for the user adding up to 12 fields in the 2nd level and 10 in the 3rd level and this was the only way I could see around it.
I had no problem retrieving the data, but any changes at the coding level were awkward and clearly the size was ridiculous.
So I am trying to avoid a similar situation this time.
Thanks
Mark
"Wolfgang Keller" <wk_at_objectarchitects.de> wrote in message
news:3B7A716D.99F6BF8_at_objectarchitects.de...
>
>
> Mark Cunningham wrote:
>
> > * number of parallel users
> > It's a single user system. The data would be stored on the user's hard
> > drive or CD.
> > * number of transactions
> > Very few - not enough to be an issue
> > * queries
> > Very simple retrieval, based on hierarchical selections.
>
> aha .. that sounds like you're better off not using any kind of database
> product ..
> Next question is the programming language you want to use because
> * e.g. in C++ and Java there are serializer libraries which will reduce
the job
>
> to almost no code or libraries that encapsulate XML ...
> * PL/I and C would be o.k. in a way that you can use recursive data
structures
> (just a plain tree ...) - you could steal this from any book on data
> structures
> using C - there are plenty of them
> * Cobol ... did it support recursion ? :-) - uuuuughh
>
> so .. we still need some more facts here ..
>
> Cheers
>
>
> Wolfgang
>
> --------------------------------------------------------------
> mailto:wk_at_objectarchitects.de
> http://www.objectarchitects.de/ObjectArchitects/
>
>
Received on Thu Aug 16 2001 - 12:46:15 CEST
