按SQL Server中的多列分组

Question

I have a table whereby some rows are identical except for a single column. I want to group by the identical rows and take the mean of the single column.

For example I have the table:

A   B   C   D
-------------
a   b   c   2
a   b   c   2
a   b   c   5
d   e   f   8
d   e   f   10

Desired result:

A   B   C   D
-------------
a   b   c   3
d   e   f   9

I could easily do this by:

SELECT A, B, C, AVG(D)
FROM [table] 
GROUP BY A, B, C

My issue is I have many columns and it is cumbersome to manually type every column to GROUP BY. For example, I want to do this but its not possible:

SELECT (* but not D), AVG(D)
FROM [table] 
GROUP BY (* but not D)

The two issues are that I don't know how to form a wildcard that excludes D, and also it seems GROUP BY does not work with wildcards.

What is the elegant way to do this?

Btw, the original table resulted by a left join that matched on non-unique values.

Answer 1

Unfortunately there isn't a magic way to do it, however it can made much easier with SSMS:

• Right click the table
• Script table as ->
• Select to ->
• New query window

This gives you have a formatted SELECT list that you can remove the aggregate columns from and copy-paste to your GROUP BY . It's at least a bit better than starting from scratch.

Answer 2

You can use dynamic SQL for this.
See working demo

DECLARE @sqlCommand nvarchar(max)
DECLARE @columnList nvarchar(max)

SET @columnList= STUFF((SELECT ',' + c.name
    from sys.columns c join sys.tables t on c.object_id =t.object_id
    and t.name ='yourtable' and c.name <>'D'
FOR XML PATH(''), Type
).value('.', 'NVARCHAR(MAX)') 
,1,1,'')
SET @sqlCommand = 'SELECT ' + @columnList + ', AVG(D) as D FROM [yourtable] GROUP BY  ' + @columnList
EXEC (@sqlCommand)

Answer 3

You can also use Select name from sys.columns where table_name = <table>

to get all the columns of the table and specify it in your Group By section

Answer 4

In SSMS , expand the Object Explorer to the desired table. Expand the table, you will see folder Column . Just drag the folder to the query window and it will list the columns of the table.