SQL Server should create clustered index before non-clustered index in exists table

Copyright: copy, please indicate the source https://blog.csdn.net/weixin_39392627/article/details/86654433

env:Windows Server 2016

       SQL Server 2016 SP2

 

There is a form to fill clustered index and non-clustered index. This is both a form and have 400 million pen data, which also has a non-clustered index.

The extent of statistical information is updated 80000000 pen.

 

1. Perform create non-clustered index before performing clustered index

command:

CREATE NONCLUSTERED INDEX [IX_TESTId_CreatOn]
ON [dbo].[TESTTB] ([TESTId],[CreatOn])
INCLUDE ([Id],[LISTId],[Amount],[Revenue],[Session])
WITH(ONLINE=ON, DATA_COMPRESSION=ON)
GO

CREATE CLUSTERED INDEX [IX_CreatOn] ON [dbo].[TESTTB]
(
[CreatOn] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = ON, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, DATA_COMPRESSION=PAGE) ON [PRIMARY]
GO

Spend a lot of time in the build clustered index, the index was mainly caused by scanning both. Use SSMS to establish cluster index will prompt other cited will rebuild.

Execution order	servername	db_date	table_name	idx_id	index_name	row_count	data_used_mb	Exec(Mins)
1	TESTDB01	TESTDB01	dbo.TESTTB	7	IX_TESTId_CreatOn	404502153	17368	51
2	TESTDB01	TESTDB01	dbo.TESTTB	1	IX_CreatOn	404502153	47358	94

 

2. The establishment of priority so that clustered index

command:

CREATE CLUSTERED INDEX [IX_CreatOn] ON [dbo].[TESTTB]
(
[CreatOn] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = ON, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, DATA_COMPRESSION=PAGE) ON [PRIMARY]
GO

CREATE NONCLUSTERED INDEX [IX_TESTId_CreatOn]
ON [dbo].[TESTTB] ([TESTId],[CreatOn])
INCLUDE ([Id],[LISTId],[Amount],[Revenue],[Session])
WITH(ONLINE=ON, DATA_COMPRESSION=ON)
GO

The establishment can be found clustered index time is considerably reduced.

Execution order	servername	db_date	table_name	idx_id	index_name	row_count	data_used_mb	Exec(Mins)
1	TESTDB01	TESTDB01	dbo.TESTTB	1	IX_CreatOn	404502153	47358	50
2	TESTDB01	TESTDB01	dbo.TESTTB	7	IX_TESTId_CreatOn	404502153	17368	43

in conclusion:

To create the index in both tables, clustered index's main priorities.

If in the table have been many non-clustered index, if you want to create an empty table and then perform data replication transfer? This is a question worth considering.