[Techtalk] Database textbook recommendations

[Wim De Smet]

On 5/14/05, Mary <mary-linuxchix at puzzling.org> wrote:
> Hey everyone,
> 
> In current projects, I'm beginning to find my database knowledge a bit
> shaky and would appreciate some pointers to reading material. I have
> perhaps unusual requirements, but perhaps not: does anyone have
> recommendations.
> 
> What I already know:
> 
>  - I know SQL, so I'm not looking for an SQL tutorial
>  - I can program in several (mainly OO) languages
>  - I have a comp sci/maths degree and while refreshers are fine, I can
>    be assumed to know something about data structures, worst/average
>    case costs, time/space tradeoff and that kind of thing. I also have some
>    knowledge of data structures, although not particularly recent or
>    practical.
> 
> What I want to know: how to apply all of this to database design and
> maintainence, how to design databases to get efficient results when
> querying datasets of varying profiles on really big databases. What
> changes when you have enormous and frequently accessed versus enormous
> and seldom accessed. This includes things like tradeoffs on updates and
> inserts versuses selects for different datasets, stress testing database
> designs, different indexing strategies and that kind of thing. MySQL and
> Postgres focus preferred at this stage.
> 
> Essentially, I want a Database Design for Computer Scientists primer, or
> Turning a Hacker into a DBA in 24 Difficult and Geeky Steps.  I'm happy
> to get recommendations that start off with stuff I do/might know and
> which get interesting in the last few chapters. Online or offline
> resources are both fine. Anyone got any?
> 
> -Mary

I'm currently studying informatics (which is kinda the same like comp
sci). We had a databases course which used:
C.J. Date, an introduction to database systems, very long amazon link:
 http://www.amazon.com/exec/obidos/tg/detail/-/0201385902/103-6048683-4044669?v=glance
It has its faults but I think it's pretty good and it also touches on
a lot of the theory. Lots of the syntax in the book is in Tutorial D,
so you won't be getting that much SQL out of it (there is a small part
of almost every chapter devoted to SQL though). It's also a very heavy
book (literally). Probably very interesting for you would be the parts
on (E)ER modelling and converting to a database. You can, btw, find
the known algorithms for that online too, but I like reading about
them in books. :-)

For our second course we're using:
R. Elmasri & B. Navathe, Fundamentals of Database Systems, long amazon link:
http://www.amazon.com/exec/obidos/tg/detail/-/0321122267/ref=pd_null_recs_b_t/103-6048683-4044669?v=glance&s=books
It sadly duplicates a lot of the information in Date's book, but also
goes deeper into the topic of object oriented databases. Personally I
think it looks a lot Date's book, but I haven't got any good reading
in it yet. Some stuff about data mining in there too which is pretty
interesting.

Anyway, I thought since my profs thought these are two good books on
the subject, they're probably not bad. Date in particular seems
obsessed with finding a theoretical sound platform for the relational
model. So the only part of databases that that book (and possibly the
second) won't highly touch on will probably be the more practical
things, like database optimalisation, which is off course what you
wanted to know about. But still, a good understanding of database
principles might be enough to come up with your own optimalisation
strategies.

HTH,
Wim