Method 1: Select Into
This is a quick and dirty method to create a temporary table. Compared to the other methods, you do not have to define the column names. Just add the ‘into #temptablename’ at the end of the columns you want selected.

select id, code
into #mytemptable
from sourcetable

--view the data
select * from #mytemptable

Method 2: Create Table
Use this method if you want more control of the definition of the table and column types

CREATE TABLE #mytemptable (
dataid int,
description varchar(30) )

-- insert some data into it
insert into #mytemptable
select id, code from sourcetable

-- view the data
select * from #mytemptable

-- drop the table if you want to
drop table #mytemptable

Method 3: Table Variable
This method is useful for programming stored procedures and user functions. Once you define the table, you can take advantage of passing table variables back and forth between code.

DECLARE @mytemptable TABLE (
dataid int,
description varchar(30) )

-- insert some data into it
insert into @mytemptable
select id, code from sourcetable

-- view the data
select * from @mytemptable

How Long are Temporary Tables Good For:
Temporary tables are good for the length of your database session. If you have a query window in management studio, once you create a temporary table, it is available until you close that query window. You can also use the ‘drop table’ statement on it.

What are Temporary Tables Good For:
– If you are analyzing data, it helps to store the results of frequently or long running sql
– If you are doing complex sql or aggregation (like or data warehouses), it might make your process simpler. I would use table variables (method 3) in that case, for performance reasons.

• Open the Options panel in the menu bar
• Navigate to Query Results -> SQL Server -> Results to Grid
• Check “Include column headers when copying or saving the results.”

SQL Server Manangement Studio Options

This works for SQL Server 2005/2008. In SQL Server 2000’s Query Analyzer there was a button in the tool bar that turned on this option.

select newid();

Creating a GUID in MYSQL:

select uuid();

Why use a GUID as a primary key?
The main reason is if you want to guarantee uniqueness between databases and servers.

I have never been a fan of GUIDs. Especially when having to write sql statements for reports. They’re difficult to scan and impossible to sort. It used to be they were non existent and then they gained popularity when trying to store objects in relational databases became a fad. Try writing reports against an object hierarchy that is stored in a relational database. Every other column is a GUID. It’s not fun.

Anyway I had one recent experience that reinforced the need of a GUID in a database. Some production data was unintentionally lost in one table(due to a roll back). This particular table was using GUID’s as a primary key. The table could not be restored since it was already out of date. Instead the table was restored to a copy of the database. This allowed a quick comparison of the two tables and a painless insert via sql. What the GUID did in this case, was guarantee sameness (instead of uniqueness).

I am still going to stick to integers as primary keys but I am now going to add an auto generated GUID column for the sake of backup/restores/import/exports/etc.