Microsoft SQL

Could you recommend to me good books and manuals on how to learn Microsoft SQL?

Yes, I could.

>>2
I doubt that, this is Microsoft we're talking about here.

>>3
No, really, I definitely could.

>>4
But you wont...right?

>>4
I would like to see a formal proof that you could in fact make such a recommendation.

>>6
I hope you don't mind if I use some axioms.

Axiom 1: If it ain't about Lisp, I can find something about it.

By Axiom 1, I could find something.

>>7
The subject in question here isnt something, the subject is something good.

At least one tip, please

>>9
Sure.
*grabs dick*
You can lick or kiss it, as long as you're a Touhou.

Try this one: leave MS and their bullshit alone and use a DBS that actually delivers: http://www.amazon.com/PostgreSQL-2nd-Edition-Korry-Douglas/dp/0672327562

If you work for a company that uses MSSQL, leave it.

If you go to school teaching MSSQL, leave it.

If you have someone recommend MSSQL, kill the fucker.

>>11
Is Postgre the best choice? I heard good things about Oracle as well, but I have to learn MSSQL due to job...and I really need the money.

SQL is dead, store and query your data as JSON.

>I heard good things about Oracle
They have excellent marketing. Oracle is proprietary, costly POS.
SQL itself is bloated, inefficient and lacks parallelism(even Oracle itself knows it).
Use open source NoSQL databases. start here: http://en.wikipedia.org/wiki/NoSQL

>>14
e/g/in quote /g/roski XD

>>12

Postgres and oracle are equivalent for most features. The greatest difference is, is that oracle has invested into commodity tools. For example a materialized view is something you have to build by hand in PostgreSQL with triggers, which changes data in the table representing the materialized view if something in the join tables is changed. In oracle this a feature, which is incorporated in the language.

Of course this will cost you some money.

For the rest is postgresql easier to use, more powerful and more expressive than MySQL. PostgreSQL follows the standard more strict, which makes a switch easier and their planner is more advanced and leads to better query plans. Especially in replication since the added hot streaming replication.

Another thing, which is interesting is that postgres can parallelize queries for you with the help of PGPool-II, which also works as a load balancer and query cache.

The last thing is that postgres has several languages to program in, just like oracle. It also enables you to create modules in C, which you can use to create new functions, types or indexes.

This makes it extremely useful.

>>14

This is a choice if your data is not strongly relational. I wouldn't use it in other settings, otherwise you are at risk reinventing  a database.

Don't use an object-relation mapper. These things don't make sense at all. The mapping between (objects, inheritance, embedded objects) and (tables,relations) is ill defined. For some constructs, there is no rational general mapping. A group by statement is difficult to model and requires the programmer to create a new object.

Some ORM's do their joins in memory. This hampers the performance of the application.

ORM are only good for the trivial cases.

Instead of this implement a small relational algebra. It is not difficult and still saves you time, but you remain in full control.

>>17

Also difficult to map is a projection. This is the fields you ask. For example:

select car_id, account_id from car_instance where id = ?

is difficult to map to an object, if not all fields are selected. But sometimes, such a query is superior. Postgres loads only the index if these two columns share a key. It doesn't have to touch the table at all, thus creating an enormous performance difference between the ORM, which needs to load for safety all the fields, and PostgreSQL.

Let a goog build your shit

>>19

Now that's lacist!