I am going to perform a table-wide update on a huge table (+200Millon records) so I am going to populate the data via SELECT into a separate table, than drop the original and rename the new one.
In one of the articles someone mentioned that it is better to create the new table with the same name in a temporary schema (e.g. 'clone') and switch it to the used one (e.g. 'dbo'), than to use the original schema with a temporary name and call sp_rename after the data is in place and the old table is dropped.
I was looking into this, but I cannot think of anything why the schema switch is better than the sp_rename. Can anyone find any good reason why is better to use the first or the second approach?
Thanks!
EDIT: I want to update the values in a specific column
EDIT2: Ultimately my question is, if I decide to go down the way of creating a new table to transfer data to which alternative to use:
CREATE TABLE dbo.newTable
...
DROP TABLE dbo.originalTable
EXEC sp_rename N'dbo.newTable', N'dbo.originalTable'
OR
CREATE TABLE clone.originalTable
...
DROP TABLE dbo.originalTable
ALTER SCHEMA dbo TRANSFER clone.originalTable
By the way, I would suggest that you WON'T populate the table by using SELECT * INTO. This will lock your source table for everyone else during the insertion, which could take quite a time.
Just a suggestion, try this instead:
SELECT TOP 0 INTO [newTable]
FROM [oldTable]
INSERT INTO [newTable]
SELECT * FROM [oldTable]
By the way, you can use sp_rename to rename your table to another name. But it won't change the schema. If you try to change the schema too it will produce a buggy table name.
You can instead try to move the table to another name. Example below:
EXEC sp_rename N'oldTable', N'oldTable_Backup'
EXEC sp_rename N'newTable', N'oldTable'
Hopefully this will help you.
Based on your edited answer the quickest way to do that is:
If you have to include default value to the column
ALTER TABLE {TABLENAME}
ADD {COLUMNNAME} {TYPE} {NULL|NOT NULL}
CONSTRAINT {CONSTRAINT_NAME} DEFAULT {DEFAULT_VALUE}
[WITH VALUES]
and then drop the old column from the table.
ALTER TABLE {TABLENAME} DROP COLUMN {OLD COLUMN}
If you have to update table column based calculated values
Disable index on the column which you are updating
Create index on the column which are in WHERE clause
Update statistics
Use WITH(NOLOCK) table hint [if you are fine with dirty read]
Update
As per edit 2, your first statement is about changing table name and second statement is about changing schema. They both are different and does not related to moving data or updating value. In this case, changing schema would be the best bet
If locking was an issue before it still is regardless of what version SQL Server that you are using. When you drop and rename you are also losing all the rights on the table.
How do I add a column after another column within MS SQL by using SQL query?
TempTable
ID int,
Type nvarchar(20),
Active bit
NewTable
ID int,
Type nvarchar(20),
Description text,
Active bit
That is what I want, how do I do that
Assuming MySQL (EDIT: posted before the SQL variant was supplied):
ALTER TABLE myTable ADD myNewColumn VARCHAR(255) AFTER myOtherColumn
The AFTER keyword tells MySQL where to place the new column. You can also use FIRST to flag the new column as the first column in the table.
It depends on what database you are using. In MySQL, you would use the "ALTER TABLE" syntax. I don't remember exactly how, but it would go something like this if you wanted to add a column called 'newcol' that was a 200 character varchar:
ALTER TABLE example ADD newCol VARCHAR(200) AFTER otherCol;
In a Firebird database the AFTER myOtherColumn does not work but you can try re-positioning the column using:
ALTER TABLE name ALTER column POSITION new_position
I guess it may work in other cases as well.
I am having problems renaming a column in SQL Server Compact Edition. I know that you can rename a table using sp_rename, but this doesn't work with columns.
I've searched for an alternative, but haven't found one.
Can I delete a column and then add a new one after a specific column? If I delete the column and add it after the a specified one the data would be lost right?
It seems that once you have created the table it can't be properly modified - is this another of the limitations of SQLCE?
It does indeed seem that SQL CE wont allow changing column names.
You're on the right track with creating a new column and deleting the old.
If you just add a column and delete the old you will lose the data so you need to issue an update statement to shift the data from the old to the new.
Something along the lines of
alter Table [dbo].[yourTable] add [newColumn] [DataType]
update yourTable set newColumn = oldColumn
alter Table [dbo].[yourTable] drop column [oldColumn]
Should create your new column, duplicate the data from old to new and then remove the old column.
Hope it helps!
sp_rename works with columns too:
EXEC sp_rename
objname = '< Table Name.Old Column Name >',
#newname = '<New Column Name>',
#objtype = 'COLUMN'
Example:
SP_RENAME 'MyTable.[MyOldColumnName]' , '[MyNewColumnName]', 'COLUMN'
UPDATE: Actually, The sp_rename procedure is not avialable in SQL CE! You can find the solution at http://www.bigresource.com/Tracker/Track-ms_sql-4Tvoiom3/
SDF Viewer has this function built in, you can also rename indexes, keys and relationships. Just right click on the name you want to change in the database treeview.
What would the command be for Microsoft SQL Server 2008 to alter an existing column to allow for more characters? Would this have an effect on any previous entries in the column if I'm only expanding it?
I have a URL column that I need to add about 100 characters to.
ALTER TABLE [table] ALTER COLUMN [column] NVARCHAR(newsize)
And increasing the size won't affect your data.
ALTER TABLE myTable ALTER COLUMN myColumn varchar(100)
GO
This would not involve the risk of losing data, because you are expanding the size of the column.
ALTER TABLE tab ALTER COLUMN c VARCHAR(200)
You can use SQL Server Management Studio to generate the code to do this yourself. Just modify the table in the designer so that the column is wider and click "Generate change script" (under the "Table designer" menu).
Depending on the change you make the code to modify a table can be fairly complex.
It is quite simple:
ALTER TABLE <table_name> MODIFY <column_name> VARCHAR(100)
How to change column order in a table using SQL query in SQL Server 2005?
I want to rearrange column order in a table using SQL query.
You cannot. The column order is just a "cosmetic" thing we humans care about - to SQL Server, it's almost always absolutely irrelevant.
What SQL Server Management Studio does in the background when you change column order there is recreating the table from scratch with a new CREATE TABLE command, copying over the data from the old table, and then dropping it.
There is no SQL command to define the column ordering.
You have to explicitly list the fields in the order you want them to be returned instead of using * for the 'default' order.
original query:
select * from foobar
returns
foo bar
--- ---
1 2
now write
select bar, foo from foobar
bar foo
--- ---
2 1
according to https://learn.microsoft.com/en-us/sql/relational-databases/tables/change-column-order-in-a-table
This task is not supported using Transact-SQL statements.
Well, it can be done, using create/ copy / drop/ rename, as answered by komma8.komma1
Or you can use SQL Server Management Studio
In Object Explorer, right-click the table with columns you want to reorder and click Design (Modify in ver. 2005 SP1 or earlier)
Select the box to the left of the column name that you want to reorder. (You can select multiple columns by holding the [shift] or
the [ctrl] keys on your keyboard.)
Drag the column(s) to another location within the table.
Then click save. This method actually drops and recreates the table, so some errors might occur.
If Change Tracking option is enabled for the database and the table, you shouldn't use this method.
If it is disabled, the Prevent saving changes that require the table re-creation option should be cleared in Tools menu > Options > Designers, otherwise "Saving changes is not permitted" error will occur.
Disabling the Prevent saving changes that require the table re-creation option is strongly advised against by Microsoft, as it leads to the existing change tracking information being deleted when the table is re-created, so you should never disable this option if Change Tracking is enabled!
Problems may also arise during primary and foreign key creation.
If any of the above errors occurs, saving fails which leaves you with the original column order.
In SQLServer Management Studio:
Tools -> Options -> Designers -> Table and Database Designers
Unselect 'Prevent saving changes that require table re-creation'.
Then:
right click the table you want to re-order the columns for.
click 'Design'.
Drag the columns to the order you want.
finally, click save.
SQLServer Management studio will drop the table and recreate it using the data.
This is similar to the question on ordering the records in the result of a query .. and typically no one likes the formally correct answer ;-)
So here it goes:
as per SQL standard, the columns in a table are not "ordered"
as a result, a select * does not force the columns to be returned in a particular order
typically, each RDBMS has a kind of "default" order (usually the order that the columns were added to the table, either in the create table' or in thealter table add ` statements
therefore, if you rely on the order of columns (because you are using the results of a query to poulate some other datastructure from the position of the columns), explicitly list the columns in the order you want them.
You can of course change the order of the columns in a sql statement. However if you want to abstract tables' physical column order, you can create a view. i.e
CREATE TABLE myTable(
a int NULL,
b varchar(50) NULL,
c datetime NULL
);
CREATE VIEW vw_myTable
AS
SELECT c, a, b
FROM myTable;
select * from myTable;
a b c
- - -
select * from vw_myTable
c a b
- - -
You can do it by creating a new table, copy all the data over, drop the old table, then renaming the new one to replace the old one.
You could also add new columns to the table, copy the column by column data over, drop the old columns, then rename new columns to match the old ones. A simple example below:
http://sqlfiddle.com/#!3/67af4/1
CREATE TABLE TestTable (
Column1 INT,
Column2 VARCHAR(255)
);
GO
insert into TestTable values(1, 'Test1');
insert into TestTable values(2, 'Test2');
GO
select * from TestTable;
GO
ALTER TABLE TestTable ADD Column2_NEW VARCHAR(255);
ALTER TABLE TestTable ADD Column1_NEW INT;
GO
update TestTable
set Column1_NEW = Column1,
Column2_NEW = Column2;
GO
ALTER TABLE TestTable DROP COLUMN Column1;
ALTER TABLE TestTable DROP COLUMN Column2;
GO
sp_rename 'TestTable.Column1_NEW', 'Column1', 'COLUMN';
GO
sp_rename 'TestTable.Column2_NEW', 'Column2', 'COLUMN';
GO
select * from TestTable;
GO
In SQLServer Management Studio:
Tools -> Options -> Designers -> Table and Database Designers
Unselect Prevent saving changes that require table re-creation.
Now you can reorder the table.
Sql server internally build the script. It create a temporary table with new changes and copy the data and drop current table then recreate the table insert from temp table. I find it from "Generate Change script" option ssms 2014. Script like this. From Here: How to change column order in a table using sql query
BEGIN TRANSACTION
SET QUOTED_IDENTIFIER ON
SET ARITHABORT ON
SET NUMERIC_ROUNDABORT OFF
SET CONCAT_NULL_YIELDS_NULL ON
SET ANSI_NULLS ON
SET ANSI_PADDING ON
SET ANSI_WARNINGS ON
COMMIT
BEGIN TRANSACTION
GO
CREATE TABLE dbo.Tmp_emps
(
id int NULL,
ename varchar(20) NULL
) ON [PRIMARY]
GO
ALTER TABLE dbo.Tmp_emps SET (LOCK_ESCALATION = TABLE)
GO
IF EXISTS(SELECT * FROM dbo.emps)
EXEC('INSERT INTO dbo.Tmp_emps (id, ename)
SELECT id, ename FROM dbo.emps WITH (HOLDLOCK TABLOCKX)')
GO
DROP TABLE dbo.emps
GO
EXECUTE sp_rename N'dbo.Tmp_emps', N'emps', 'OBJECT'
GO
COMMIT
If your table has enough columns then you can try this. First create a new table with preferred order of columns.
create table new as select column1,column2,column3,....columnN from table_name;
Now drop the table using drop command
drop table table_name;
now rename the newly created table to your old table name.
rename new to table_name;
now select the table, you have your columns rearranged as you preferred before.
select * from table_name;
Not sure if still relevant, but SSMS can generate a change scripts for this.
Re-order (drag the column) the table in Designer View
Click on 'Generate Change Script'
The generated script contains the script which does the following:
Create a temporary table
Adds the constraints, relationships and triggers from original table to temporary table
Drop original table
Rename temporary table to original table name
If you have not yet added any data into your table yet, there is one way to move the columns around.
Try this:
In SSMS, click Tools > Options > Designers > Table and Database Designers > Uncheck the box next to Prevent saving changes that require table re-creation > Click OK.
In the object tree, right-click on your table and select Design > in the thin column to the left of the Column Name column, you can click and drag the columns around to wherever you want them. When you're done, just go to close the Design tab and SSMS will ask you if you want to save your changes, click OK.
Optional:
3. Re-enable the checkbox for the option from Step 1 to re-secure your table.
Hope this helps someone!
Credit goes to Microsoft:
https://learn.microsoft.com/en-us/troubleshoot/sql/ssms/error-when-you-save-table#more-information
At the end of the day, you simply cannot do this in MS SQL. I recently created tables on the go (application startup) using a stored Procedure that reads from a lookup table. When I created a view that combined these with another table I had manually created earlier one (same schema, with data), It failed - simply because I was using ''Select * UNION Select * ' for the view. At the same time, if I use only those created through the stored procedure, I am successful.
In conclusion: If there is any application which depends on the order of column it is really not good programming and will for sure create problems in the future. Columns should 'feel' free to be anywhere and be used for any data process (INSERT, UPDATE, SELECT).
You can achieve it with these steps:
remove all foreign keys and primary key of the original table.
rename the original table.
using CTAS create the original table in the order you want.
drop the old table.
apply all constraints back to the original table
If the columns to be reordered have recently been created and are empty, then the columns can be deleted and re-added in the correct order.
This happened to me, extending a database manually to add new functionality, and I had missed a column out, and when I added it, the sequence was incorrect.
After finding no adequate solution here I simply corrected the table using the following kind of commands.
ALTER TABLE tablename DROP COLUMN columnname;
ALTER TABLE tablename ADD columnname columntype;
Note: only do this if you don't have data in the columns you are dropping.
People have said that column order does not matter. I regularly use SQL Server Management Studio "generate scripts" to create a text version of a database's schema. To effectively version control these scripts (git) and to compare them (WinMerge), it is imperative that the output from compatible databases is the same, and the differences highlighted are genuine database differences.
Column order does matter; but just to some people, not to everyone!
Use
SELECT * FROM TABLE1
which displays the default column order of the table.
If you want to change the order of the columns.
Specify the column name to display correspondingly
SELECT COLUMN1, COLUMN5, COLUMN4, COLUMN3, COULMN2 FROM TABLE1
you can use indexing.. After indexing, if select * from XXXX results should be as per the index, But only result set.. not structrue of Table
In order to have a specific column order You need to select column by column in the order You wish.
Selection order dictates how columns will be ordered in output.
Try this command:
alter table students modify age int(5) first;
This will change the position of age to the first position.
You can change this using SQL query. Here is sql query to change the sequence of column.
ALTER TABLE table name
CHANGE COLUMN `column1` `column1` INT(11) NOT NULL COMMENT '' AFTER `column2`;
alter table name modify columnname int(5) first;
will bring the column to first
alter table name modify columnname int(5) after (tablename);
This worked for me on Oracle DB:
select column1, column2, t.* from table t
Example: Change position of field_priority after field_price in table status.
ALTER TABLE `status` CHANGE `priority` `priority` INT(11) NULL DEFAULT NULL AFTER `price`;