在视图或表上显示索引？

Question

Can anyone help me with this one.

Where should I put the index for this one since it is a view? Does indexing the sub query (table) help optimize this, or should I just put an index on the view instead?

Thank you very much.

1. INDEX on vw_fact_test ?
2. INDEX on A1LE_Project ?

Code:

ALTER VIEW [dbo].[vw_fact_test]
AS
  SELECT     
      A.Column1
      A.Column2, 
      A.Column3, 
      CASE 
         WHEN Column4 IS NULL OR Column4 = '' OR Column4 = 'NULL' 
           THEN 'unknown' 
           ELSE Column4 
      END AS Column4, 
      CASE 
        WHEN Column5 IS NULL OR Column5 = '' OR Column5 = 'NULL' 
          THEN 'unknown' 
          ELSE Column5 
      END AS Column5, 
      A.Column6, 
      dbo.M3Ofc_Lookups.Column7
FROM         
    (SELECT     
        Column1, 
        Column2, 
        Column3, 
        Column4, 
        Column5, 
        Column6, 
    FROM 
        dbo.A1LE_Project) AS A 
INNER JOIN 
    dbo.M3Ofc_Lookups ON A.[Office Code] = dbo.M3Ofc_Lookups.[Office Code] 
WHERE 
    (A.Column2 <> 0) OR
    (A.Column3 <> 0) OR
    (A.Column4 <> 0)

Answer 1

Firstly you can simplify your select statement, which makes it easier to read and think about the indexing. You don't need a subquery.

SELECT A.Column1, A.Column2, A.Column3, 
        CASE 
            WHEN Column4 IS NULL OR Column4 = '' THEN 'unknown' 
            ELSE Column4 
        END AS Column4, 
        CASE 
            WHEN Column5 IS NULL OR Column5 = '' THEN 'unknown' 
            ELSE Column5 
        END AS Column5, 
        A.Column6, B.Column7
    FROM dbo.A1LE_Project AS A 
        INNER JOIN dbo.M3Ofc_Lookups AS B
            ON A.[Office Code] = B.[Office Code] 
    WHERE (A.Column2 <> 0) OR (A.Column3 <> 0) OR (A.Column4 <> 0);

You need to consider the sizes of the tables and where you want to improve the performance. You also need to examine the execution plan.

If table B is a lookup table, it is likely to be fairly small. So the query processor could do a table scan with little cost. If it is sizeable, making [Office Code] the clustered primary key would help. Or possibly putting an index on [Office Code] and Column7.

How many records are in table A, and what are the types and sizes of columns?

If table A has a large number of records, and the select statement will yield a small proportion, then putting indexes on Column2, Column3, and Column4 might help. The query processor would be able to quickly select early on just the records it needs. You would need to test this, and see the effect on the execution plan.

If columns 1 to 6 are small, and the table contains a lot of other columns, then adding a covering index could help.

Before adding any indexes you need to consider how the database is being updated and the impact on performance. Is the database only being used for reporting, or is it being updated as well?

Adding an index on the view itself could have a severe impact on update performance. It's almost like creating an additional table which is kept automatically in step. See SQL Server Books Online > Create Indexed Views. "The first index created on a view must be a unique clustered index. After the unique clustered index has been created, you can create additional nonclustered indexes. Creating a unique clustered index on a view improves query performance because the view is stored in the database in the same way a table with a clustered index is stored."