One of the new T-SQL features in SQL 2012 is the Sequence object.
The object is used to generate identity values for a table. One advantage is that a value can be generated without having to first make an insert into a table.
To create:
create sequence dbo.TestSequence
as int
start with 1
increment by 1
cycle
cache 10
So we can specify the value the sequence starts with, and by how much it gets incremented. We can specify CYCLE to reuse values once we get to the largest value allowed by the data type, or NO CYCLE to raise an exception once the limit is reached.
We can also specify CACHE to obtain a specified number of values at once.
There are two ways to use the Sequence. We can go to the Sequence on each insert to get the next value, or we can use the Sequence as part of the table definition.
— Not part of table definition
create table TestTable1 (
RecordNumber int not null primary key clustered,
DisplayName varchar(25) not null
)
go
insert into TestTable1 (RecordNumber, DisplayName)
values (next value for dbo.TestSequence, ‘Record One’)
insert into TestTable1 (RecordNumber, DisplayName)
values (next value for dbo.TestSequence, ‘Record Two’)
select * from TestTable1
go
— Part of table definition
create table TestTable2 (
RecordNumber int not null primary key clustered default(next value for dbo.TestSequence),
DisplayName varchar(25) not null
)
go
insert into TestTable2 (DisplayName)
values (‘Record Three’)
select * from TestTable2
go
We can query the catalog view to see all Sequences:
select * from sys.sequences
We can also use ALTER SEQUENCE to change the Sequence, and DROP SEQUENCE to remove it from the database.
SQL Rob 