SQL server identity column values start at 0 instead of 1

Question

I've got a strange situation with some tables in my database starting its IDs from 0, even though TABLE CREATE has IDENTITY(1,1). This is so for some tables, but not for others. It has worked until today.

I've tried resetting identity column:

DBCC CHECKIDENT (SyncSession, reseed, 0);

But new records start with 0. I have tried doing this for all tables, but some still start from 0 and some from 1.

Any pointers?

(i'm using SQL Server Express 2005 with Advanced Services)

Answer 1

From DBCC CHECKIDENT

DBCC CHECKIDENT ( table_name, RESEED, new_reseed_value )

If no rows have been inserted to the table since it was created, or all rows have been removed by using the TRUNCATE TABLE statement, the first row inserted after you run DBCC CHECKIDENT uses new_reseed_value as the identity. Otherwise, the next row inserted uses new_reseed_value + the current increment value.

So, this is expected for an empty or truncated table.

Answer 2

If you pass a reseed value the DB will start the identity from that new value:

DBCC CHECKIDENT (SyncSession, RESEED, 0); --next record should be 0 + increment

You don't have to pass the a value though, if you don't IDENTITY(a,b) will be used instead:

DBCC CHECKIDENT (SyncSession, RESEED); --next record should be the seed value 'a'

This is usually better practice, as it leaves the table closer to its initial created state.

Answer 3

This is logical, since you've changed (reseeded) the identity value to zero ?

DBCC CHECKIDENT (SyncSession, reseed, 1)

will reseed your identity column, and make sure that the first new record will start with 1.

Answer 4

I have the same problem, restoring from a backup after modifying the DB. I just add a dummy record and then delete it... then set RESEED to 0. Seems to work.

Answer 5

The currently accepted answer only explains this annoying phenomenon. Only one answer offers some sort of a solution, but not really practical because it requires a dummy insertion, which makes it hard to generalize.

The only generic solution is to reseed the identity value, then check the current identity value and reseed it again when it's 0. This can be done by a stored procedure:

CREATE OR ALTER PROCEDURE ReseedIdentity
    @tableName SYSNAME
AS
BEGIN
    DBCC CHECKIDENT(@tableName, RESEED, 0)
    IF IDENT_CURRENT(@tableName) = 0
    BEGIN
        DBCC CHECKIDENT(@tableName, RESEED, 1)
    END
END

This will always start new records at identity value 1, whether it's a new table, after truncating or after deleting all records.

If there are identity specifications starting at higher seed values a somewhat more advanced version can be used, which is a generalization of the former:

CREATE OR ALTER PROCEDURE ReseedIdentity
    @tableName SYSNAME
AS
BEGIN
    DECLARE @seed NUMERIC(18,0) = IDENT_SEED(@tableName) - 1;
    DBCC CHECKIDENT(@tableName, RESEED, @seed)
    IF IDENT_CURRENT(@tableName) = @seed
    BEGIN
        SET @seed = @seed + 1
        DBCC CHECKIDENT(@tableName, RESEED, @seed)
    END
END

Answer 6

Try this

DECLARE @c TABLE (TanvtechId varchar(10),NewTanvtechId Varchar(10))
INSERT INTO @c
SELECT TanvtechId , Row_Number() OVER (ORDER BY TanvtechId ) from Tanvtech 

UPDATE G
SET G.TanvtechId =a.NewTanvtechId 
FROM Tanvtech as G INNER JOIN @c as a ON a.TanvtechId =G.TanvtechId 

Answer 7

DBCC CHECKIDENT ( Table_Name, RESEED, 0 )

This is a way to start an id with Zero(0), then delete all the rows from table and again put the data back into the table.