SQL Count(*) VS Count(1)
Download as PPTX, PDF0 likes380 views
This document discusses using indexes in SQL to improve query performance. It shows how creating a nonclustered index on columns in a table can change a query from doing a full table scan to an index scan, using fewer I/O resources. It also examines how different indexes may perform differently for the same query based on the column indexed. The document demonstrates checking index fragmentation and shows how a uniqueidentifier primary key can become fragmented over time with multiple inserts.
![SELECT a.[index_id], a.[page_count], b.name AS index_name
FROM sys.dm_db_index_physical_stats(
DB_ID (),
OBJECT_ID ('CTest'),
NULL,
NULL,
'LIMITED') a
INNER JOIN sys.indexes b
ON a.object_id = b.object_id AND a.index_id = b.index_id;
13](https://image.slidesharecdn.com/sql-161229174418/85/SQL-Count-VS-Count-1-13-320.jpg)
SQL Count(*) VS Count(1)
- 1. 1
- 2. SELECT COUNT (*) FROM mytable 一定會使用Table Scan? 2
- 3. 3
- 4. 4
- 5. 5
- 6. CREATETABLE CTest (c1 INT IDENTITY , c2 BIGINT DEFAULT 1 , c3 CHAR (1000) DEFAULT 'a' ); GO SET NOCOUNT ON; GO INSERT INTO CTest DEFAULTVALUES; 6
- 7. Enable execution plan viewers in SSMS Include Actual Execution Plan 7
- 8. SELECT COUNT (*) from CTest; 8
- 9. CREATE NONCLUSTERED INDEX CTest_1 ON CTest (c2); 9
- 10. SELECT COUNT (*) from CTest; 變成了Index Scan 因為使用 CTest_1 index 會比使用Table Scan 使用更少的 i/o 10
- 11. CREATE NONCLUSTERED INDEX CTest_2 ON CTest (c1); 11
- 12. SELECT COUNT (*) from CTest; 跟 CTest_1 Index 相比 CTest_2 Index 使用更少的 i/o 12
- 13. SELECT a.[index_id], a.[page_count], b.name AS index_name FROM sys.dm_db_index_physical_stats( DB_ID (), OBJECT_ID ('CTest'), NULL, NULL, 'LIMITED') a INNER JOIN sys.indexes b ON a.object_id = b.object_id AND a.index_id = b.index_id; 13
- 14. 14
- 16. 16
- 18. 18
- 20. 20
- 21. create table tblSAMPLE_1 (xnewid uniqueidentifier default(newid()) primary key, xname varchar(10) ); go insert into tblSAMPLE_1(xname) values('rainmaker') go 500 21
- 22. 22
- 23. 23
- 24. SELECT a.index_id, name, avg_fragmentation_in_percent FROM sys.dm_db_index_physical_stats (DB_ID(N'tempdb'), OBJECT_ID(N'tblSAMPLE_1'), NULL, NULL, NULL) AS a JOIN sys.indexesAS b ON a.object_id = b.object_id AND a.index_id = b.index_id; 24
Editor's Notes
- #3: http://www.sqlskills.com/blogs/paul/which-index-will-sql-server-use-to-count-all-rows/