December 6, 2012 at 2:25 am
I have a table which has some n no of rows. The table has Key, value1, value2, value3, etc fields. Now i need to insert a row which should be derived from the existing rows.
eg: below are the data in the table
KeyValue1Value2Value3
ky01010015090
ky02015020010
ky03058040010
ky04016010010
ky05011025010
ky06019028010
ky07035027010
I need to insert a row as below
ky05531050-10
the Value1, 2 and 3 are derived from the rows ky030 - ky040 - ky050
If it addition i can simply write
SELECT'ky055', SUM(Value1), SUM(Value2), SUM(Value3)
FROM table_name
WHERE Key in (ky030,ky040,ky050)
but i need to subtract the rows...
Can anyone help me to find a query to generate the value
December 6, 2012 at 9:06 am
Please read this.
http://www.sqlservercentral.com/articles/Best+Practices/61537/[/url]
Here is a query that will make THIS specific example work.
It rely's on your Key column to be ordered correctly.
If the order of your keys is guaranteed then fine, but I would seriously look at the design of this table and data storage strategy before worrying about creating these queries.
DECLARE @Table TABLE ( Varchar(10),Value1 INT, Value2 INT, Value3 INT)
INSERT INTO @Table
SELECT 'ky010', 100, 150, 90 UNION ALL
SELECT 'ky020', 150, 200, 10 UNION ALL
SELECT 'ky030', 580, 400, 10 UNION ALL
SELECT 'ky040', 160, 100, 10 UNION ALL
SELECT 'ky050', 110, 250, 10 UNION ALL
SELECT 'ky060', 190, 280, 10 UNION ALL
SELECT 'ky070', 350, 270, 10
SELECT 'ky055',
SUM(CASE WHEN ID =1 THEN Value1 ELSE Value1 * -1 END),
SUM(CASE WHEN ID =1 THEN Value2 ELSE Value2 * -1 END),
SUM(CASE WHEN ID =1 THEN Value3 ELSE Value3 * -1 END)
FROM (
SELECT RANK() OVER(ORDER BY [Key]) AS id, [Key], Value1, Value2, Value3
FROM @table
WHERE [Key] in ('ky030','ky040','ky050')
)t
December 9, 2012 at 9:39 pm
Thanks Ray,
Its working...
Also the article is nice once. I will follow that....
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply