[Cialug] OT?: DBMs

[Jeffrey C. Ollie]

On Mon, 2007-10-29 at 11:16 -0500, Stuart Thiessen wrote:
> I am curious why?  It was my understanding that if you have equal  
> parts of reading and writing, then SQL databases were more recommended  
> because the speed of handling transactions both ways and handling  
> multiple writes from multiple sources. On the other hand, if you were  
> dealing with more reading and only occasional writing (usually from a  
> managed source) then a DBM or XML approach were better because the  
> "increased complexity" of SQL didn't provide additional benefits over  
> DBM or XML. The ability of DBM and XML to store the data like your  
> application views it provided a better benefit in terms of speed and  
> processing than add the necessary processing to convert your data  
> between the application's processing of it and the relational database  
> structure. At least that was what my reading so far has indicated.

Databases like GDBM or Berkeley DB don't support multiple indexes per
table or joining tables, so for anything more complex than a single
index on a single table you're reduced to linear searching of the tables
and writing the code to join tables manually.  SQLite is probably fine
for prototyping, but I wouldn't run anything I considered "production"
with it unless I had no other choice.  I only run MySQL if software that
I need to run won't run on PostgreSQL.  Any projects that I develop
myself I use PostgreSQL.  And even for prototyping/development I'd
rather run on the database that I'm ultimately going to be using in
production.

With any database there's going to be some overhead of converting
between the format that you have in-memory and the format that the
database stores.

Jeff
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://cialug.org/pipermail/cialug/attachments/20071029/a007715a/attachment.pgp