I am getting error for the below code
ALTER TABLE ADM_Roles ALTER COLUMN RoleID int IDENTITY (1, 1)
Incorrect syntax near the keyword IDENTITY.
You cannot change an existing column into an IDENTITY column - you will need to add a new column that has the identity flag:
ALTER TABLE dbo.ADM_Roles
ADD NewRoleID INT IDENTITY (1, 1)
If you need to, you can later on drop the old column and rename the new column to the old name:
ALTER TABLE dbo.ADM_Roles DROP COLUMN RoleID
EXEC sp_rename #objName = 'dbo.ADM_Roles.NewRoleID',
#newName = 'RoleID',
#objType = 'COLUMN'
From the MSDN
You can't alter the existing columns for identity.
You have 2 options,
Create a new table with identity & drop the existing table
Create a new column with identity & drop the existing column But
take special care when these columns have any constraints / relations.
Example approach:
In this approach you can’t retain the existing data values on the newly created identity column;
The identity column will hold the sequence of number
Alter Table Names Add Id_new Int Identity(1,1)
Go
Alter Table Names Drop Column ID
Go
Exec sp_rename 'Names.Id_new', 'ID','Column'
you have to remove the word "int".
ALTER TABLE ADM_Roles ALTER COLUMN RoleId IDENTITY (1, 1);
Related
I'm trying to add an auto increment column in an existing table of DB2.
DB2 version is v10.5.0.5.
Following is my query:
alter table DB2INST1.AAA_BJ_BOND
ADD COLUMN id INTEGER NOT NULL DEFAULT 0;
ALTER TABLE DB2INST1.AAA_BJ_BOND ALTER COLUMN id
set generated always as identity (start with 1);
but I got following error:
"com.ibm.db2.jcc.am.SqlSyntaxErrorException: ALTER TABLE "DB2INST1.AAA_BJ_BOND"
specified attributes for column "ID" that are not compatible with the existing
column.. SQLCODE=-190, SQLSTATE=42837, DRIVER=4.13.127"
What can I do to solve this problem?
You must drop the column DEFAULT value first.
This is mentioned in the description of SQL0190N:
If SET GENERATED ALWAYS AS (expression) is specified, but the column
is already defined with a form of generation (default, identity, or
expression) and there is no corresponding DROP in the same statement.
ALTER TABLE DB2INST1.AAA_BJ_BOND
ALTER COLUMN id drop default;
ALTER TABLE DB2INST1.AAA_BJ_BOND ALTER COLUMN id
set generated always as identity (start with 1);
Now I have successfully added auto-increasing ID to the table through the following three steps:
ALTER TABLE DB2INST1.AAA_SEAT ADD COLUMN ID INTEGER NOT NULL DEFAULT 0;
ALTER TABLE DB2INST1.AAA_SEAT ALTER COLUMN ID DROP DEFAULT;
ALTER TABLE DB2INST1.AAA_SEAT ALTER COLUMN ID SET GENERATED ALWAYS AS IDENTITY (START WITH 1);
I have a table with following schema in my DB2 database.
CREATE TABLE IDN_OAUTH_CONSUMER_APPS (
CONSUMER_KEY VARCHAR (255) NOT NULL,
CONSUMER_SECRET VARCHAR (512),
USERNAME VARCHAR (255),
TENANT_ID INTEGER DEFAULT 0,
APP_NAME VARCHAR (255),
OAUTH_VERSION VARCHAR (128),
CALLBACK_URL VARCHAR (1024),
GRANT_TYPES VARCHAR (1024)
/
I need to add a new column ID of Type integer not null auto increment, and make it the primary key. How can I do that without deleting the table?
I could do this successfully using following set of queries.
ALTER TABLE IDN_OAUTH_CONSUMER_APPS ADD COLUMN ID INTEGER NOT NULL DEFAULT 0
CREATE SEQUENCE IDN_OAUTH_CONSUMER_APPS_SEQUENCE START WITH 1 INCREMENT BY 1 NOCACHE
CREATE TRIGGER IDN_OAUTH_CONSUMER_APPS_TRIGGER NO CASCADE BEFORE INSERT ON IDN_OAUTH_CONSUMER_APPS REFERENCING NEW AS NEW FOR EACH ROW MODE DB2SQL BEGIN ATOMIC SET (NEW.ID) = (NEXTVAL FOR IDN_OAUTH_CONSUMER_APPS_SEQUENCE); END
REORG TABLE IDN_OAUTH_CONSUMER_APPS
UPDATE IDN_OAUTH_CONSUMER_APPS SET ID = IDN_OAUTH_CONSUMER_APPS_SEQUENCE.NEXTVAL
And then add primary key using alter table.
Use a multi-step approach:
add the column ALTER TABLE ADD... with just the integer data type and as nullable
update the table to set the intended identity values for that column
alter the table to add the auto-generation
alter the table to add the primary key on that column
You need to have multiple steps because the identity values need to be added manually. Syntax and examples for ALTER TABLE can be found here.
There is an easy way to do it. Just run the alters above:
ALTER TABLE idn_oauth_consumer_apps ADD COLUMN id INTEGER NOT NULL DEFAULT 0;
ALTER TABLE idn_oauth_consumer_apps ALTER COLUMN id SET GENERATED ALWAYS AS IDENTITY;
It is simple and fast even on big tables. Tested and working on DB2 for i V7R2.
I recommend using this approach. It does not require creating any satellite objects - no triggers, sequences, etc...
alter table test.test2 add column id integer not null default 0;
alter table test.test2 alter column id drop default;
alter table test.test2 alter column id set generated always as identity;
call sysproc.admin_cmd ('reorg table test.test2');
update test.test2 set id = default;
commit;
If using "db2" cli then the reorg command may be run directly without the "call sysproc.admin_cmd" wrapper.
Create a new table with the primary key field. Insert the records from the old table. Drop the old table and if you can, rename the new one. If you can't rename it, recreate it and populate from the one that now has the records.
Building on Chamila Wijayarathna's answer, I used the following:
ALTER TABLE IDN_OAUTH_CONSUMER_APPS ADD COLUMN ID INTEGER NOT NULL DEFAULT 0
CREATE SEQUENCE IDN_OAUTH_CONSUMER_APPS_ID_SEQUENCE START WITH 1 INCREMENT BY 1 NOCACHE
CREATE TRIGGER IDN_OAUTH_CONSUMER_APPS_ID_TRIGGER NO CASCADE BEFORE INSERT ON
IDN_OAUTH_CONSUMER_APPS REFERENCING NEW AS NEW
FOR EACH ROW MODE DB2SQL BEGIN ATOMIC SET (NEW.ID) = (NEXTVAL FOR
IDN_OAUTH_CONSUMER_APPS_ID_SEQUENCE); END
REORG TABLE IDN_OAUTH_CONSUMER_APPS
UPDATE IDN_OAUTH_CONSUMER_APPS SET ID = IDN_OAUTH_CONSUMER_APPS_ID_SEQUENCE.NEXTVAL
ALTER TABLE IDN_OAUTH_CONSUMER_APPS ADD PRIMARY KEY (ID)
REORG TABLE IDN_OAUTH_CONSUMER_APPS
Then to reverse:
REORG TABLE IDN_OAUTH_CONSUMER_APPS
ALTER TABLE IDN_OAUTH_CONSUMER_APPS DROP PRIMARY KEY
DROP TRIGGER IDN_OAUTH_CONSUMER_APPS_ID_TRIGGER
DROP SEQUENCE IDN_OAUTH_CONSUMER_APPS_ID_SEQUENCE
ALTER TABLE IDN_OAUTH_CONSUMER_APPS DROP COLUMN ID
REORG TABLE IDN_OAUTH_CONSUMER_APPS
Tried this on DB2 for z/OS v12 and it worked:
alter table TABLE_NAME add column id integer generated always as identity
I have multiple tables in database all table has id column which is primary key.
I want a script by which i can add identity property to all tables at once rather than I go and change one by one.
You can't alter the existing columns for identity.
You have 2 options:
Create a new table with identity & drop the existing table
Create a new column with identity & drop the existing column
But take spl care when these columns have any constraints / relations.
For already craeted table Names
Drop table Names
Create table Names
(
ID int,
Name varchar(50)
)
Insert Into Names Values(1,'SQL Server')
Insert Into Names Values(2,'ASP.NET')
Insert Into Names Values(4,'C#')
In this Approach you can retain the existing data values on the newly
created identity column
CREATE TABLE dbo.Tmp_Names
(
Id int NOT NULL IDENTITY (1, 1),
Name varchar(50) NULL
) ON [PRIMARY]
go
SET IDENTITY_INSERT dbo.Tmp_Names ON
go
IF EXISTS(SELECT * FROM dbo.Names)
INSERT INTO dbo.Tmp_Names (Id, Name)
SELECT Id, Name FROM dbo.Names TABLOCKX
go
SET IDENTITY_INSERT dbo.Tmp_Names OFF
go
DROP TABLE dbo.Names
go
Exec sp_rename 'Tmp_Names', 'Names'
In this approach you can’t retain the existing data values on the
newly created identity column;
The identity column will hold the sequence of number
Alter Table Names Add Id_new Int Identity(1,1)
Go
Alter Table Names Drop Column ID
Go
Exec sp_rename 'Names.Id_new', 'ID','Column'
Source 1
What you can do is write a quick query to generate the SQL for you
like so:
USE INFORMATION_SCHEMA;
SELECT
CONCAT("ALTER TABLE `", TABLE_SCHEMA,"`.`", TABLE_NAME, "` CONVERT TO CHARACTER SET UTF8;")
AS MySQLCMD FROM TABLES
WHERE TABLE_SCHEMA = "your_schema_goes_here";
Then you can run the output from this to do what you need.
Source 2
EDIT
You could check Altering Multiple Tables at once
not working
ALTER TABLE ProductInProduct ALTER COLUMN Id KEY IDENTITY (1, 1);
Check Image
I have a table ProductInProduct is want its id should be Unique..
You cannot "convert" an existing column into an IDENTITY column - you will have to create a new column as INT IDENTITY:
ALTER TABLE ProductInProduct
ADD NewId INT IDENTITY (1, 1);
Update:
OK, so there is a way of converting an existing column to IDENTITY. If you absolutely need this - check out this response by Martin Smith with all the gory details.
You can't alter the existing columns for identity.
You have 2 options,
Create a new table with identity & drop the existing table
Create a new column with identity & drop the existing column
Approach 1. (New table) Here you can retain the existing data values on the newly created identity column.
CREATE TABLE dbo.Tmp_Names
(
Id int NOT NULL
IDENTITY(1, 1),
Name varchar(50) NULL
)
ON [PRIMARY]
go
SET IDENTITY_INSERT dbo.Tmp_Names ON
go
IF EXISTS ( SELECT *
FROM dbo.Names )
INSERT INTO dbo.Tmp_Names ( Id, Name )
SELECT Id,
Name
FROM dbo.Names TABLOCKX
go
SET IDENTITY_INSERT dbo.Tmp_Names OFF
go
DROP TABLE dbo.Names
go
Exec sp_rename 'Tmp_Names', 'Names'
Approach 2 (New column) You can’t retain the existing data values on the newly created identity column, The identity column will hold the sequence of number.
Alter Table Names
Add Id_new Int Identity(1, 1)
Go
Alter Table Names Drop Column ID
Go
Exec sp_rename 'Names.Id_new', 'ID', 'Column'
See the following Microsoft SQL Server Forum post for more details:
http://social.msdn.microsoft.com/forums/en-US/transactsql/thread/04d69ee6-d4f5-4f8f-a115-d89f7bcbc032
You don't set value to default in a table. You should clear the option "Default value or Binding" first.
I have a table with column ID that is identity one. Next I create new non-identity column new_ID and update it with values from ID column + 1. Like this:
new_ID = ID + 1
Next I drop ID column and rename new_ID to name 'ID'.
And how to set Identity on this new column 'ID'?
I would like to do this programmatically!
As far as I know, you have to create a temporary table with the ID field created as IDENTITY, then copy all the data from the original table. Finally, you drop the original table and rename the temporary one. This is an example with a table (named TestTable) that contains only one field, called ID (integer, non IDENTITY):
BEGIN TRANSACTION
GO
CREATE TABLE dbo.Tmp_TestTable
(
ID int NOT NULL IDENTITY (1, 1)
) ON [PRIMARY]
GO
SET IDENTITY_INSERT dbo.Tmp_TestTable ON
GO
IF EXISTS(SELECT * FROM dbo.TestTable)
EXEC('INSERT INTO dbo.Tmp_TestTable (ID)
SELECT ID FROM dbo.TestTable WITH (HOLDLOCK TABLOCKX)')
GO
SET IDENTITY_INSERT dbo.Tmp_TestTable OFF
GO
DROP TABLE dbo.TestTable
GO
EXECUTE sp_rename N'dbo.Tmp_TestTable', N'TestTable', 'OBJECT'
GO
COMMIT
Looks like SQL Mobile supports altering a columns identity but on SQL Server 2005 didn't like the example from BOL.
So your options are to create a new temporary table with the identity column, then turn Identity Insert on:
Create Table Tmp_MyTable ( Id int identity....)
SET IDENTITY_INSERT dbo.Tmp_Category ON
INSERT Into Tmp_MyTable (...)
Select From MyTable ....
Drop Table myTable
EXECUTE sp_rename N'dbo.Tmp_MyTable', N'MyTable', 'OBJECT'
Additionally you can try and add the column as an identity column in the first place and then turn identity insert on. Then drop the original column. But I am not sure if this will work.
Guessing you didn't have much luck with your task....
In table design, you should be able to go into properties and under Identity Specification change (Is Identity) to Yes and assign the column primary key if it formerly had the primary key.
From SqlServerCentral.com
Changing from Non-IDENTITY to IDENTITY and vice versa
An Identity is a property that is set at the time the table is created or a new column is added in alter table statement. You can't alter the column and set it to identity and it is impossible to have two identity columns within the same table.
Depending on the size of the table, is it possible to simply create a new table? copy over the schema of the old one and then use SET IDENTITY_INSERT ON to populate the new identity column with what you want from the old one.