another way to write this sql - sql-server

Question

Environment: Sql Server 2008
have rows of a column that contains comma separated values.
what to get the row even if a single product exists in that csv.
this is how i can do it but was just wondering if another way to write it?
SELECT * FROM REWARDS
WHERE ProductCsv like '%banana%'
or ProductCsv like '%strawberry%'
or ProductCsv like '%orange%'

Answer 1

Your current query doesn't seem to accurately capture the results you want. What if we have a row like this:
bananaberry cream pie,strawberry shortcake,orange juice
Should this match your query or not? I think this would be more accurate:
WHERE ',' + ProductCsv + ',' LIKE '%,banana,%'
OR ',' + ProductCsv + ',' LIKE '%,strawberry,%'
OR ',' + ProductCsv + ',' LIKE '%,orange,%'
If you're just trying to find one item, this is probably much more efficient:
WHERE ',' + ProductCsv + ',' LIKE '%,banana,%'
You probably want to use a split function. For example:
CREATE FUNCTION dbo.SplitStrings_XML
(
#List NVARCHAR(MAX),
#Delim NVARCHAR(255)
)
RETURNS TABLE
WITH SCHEMABINDING
AS
RETURN
(
SELECT Item = y.i.value('(./text())[1]', 'nvarchar(4000)')
FROM
(
SELECT x = CONVERT(XML, '<i>' + REPLACE(#List, #Delim, '</i><i>')
+ '</i>').query('.')
) AS a CROSS APPLY x.nodes('i') AS y(i)
);
GO
So now you can say:
SELECT * FROM dbo.REWARDS AS r
WHERE EXISTS
(
SELECT 1 FROM dbo.REWARDS AS r2
CROSS APPLY dbo.SplitStrings_XML(r2.ProductCsv, ',') AS x
WHERE x.Item IN ('orange','strawberry','banana')
AND r2.[key] = r.[key]
);
Or more simply:
SELECT DISTINCT r.ProductCsv --, other columns
FROM dbo.REWARDS AS r
CROSS APPLY dbo.SplitStrings_XML(r.ProductCsv, ',') AS x
WHERE x.Item IN ('orange','strawberry','banana');
The XML approach is a little brittle depending on the kinds of strings can be stored in the table, but there are many alternatives (including passing in your set via a TVP instead of as a list or separate values). For some much deeper background on split functions see these blog posts:
http://www.sqlperformance.com/2012/07/t-sql-queries/split-strings
http://www.sqlperformance.com/2012/08/t-sql-queries/splitting-strings-follow-up
http://www.sqlperformance.com/2012/08/t-sql-queries/splitting-strings-now-with-less-t-sql
That all said, I don't know if this is any better than what you have now.

Answer 2

Have a look at the following example using XML
DECLARE #Table TABLE(
CommaList VARCHAR(MAX)
)
INSERT INTO #Table SELECT 'banana,test,hello'
INSERT INTO #Table SELECT 'homer,banana,test,hello'
INSERT INTO #Table SELECT 'homer,banana'
INSERT INTO #Table SELECT '1,2,3'
;WITH XMLValues AS (
SELECT *,
CAST('<d>' + REPLACE(CommaList, ',', '</d><d>') + '</d>' AS XML) XMLValue
FROM #Table t
)
, SplitValues AS (
SELECT xv.*,
T2.Loc.value('.','VARCHAR(MAX)') SplitValue
FROM XMLValues xv
CROSS APPLY XMLValue.nodes('/d') as T2(Loc)
)
SELECT DISTINCT
CommaList
FROM SplitValues
WHERE SplitValue = 'banana'
xml Data Type Methods
nodes() Method (xml Data Type)
value() Method (xml Data Type)
Using Common Table Expressions

Answer 3

You could store all the values to be compared in a table
DECLARE #Product_list TABLE(
products VARCHAR(50)
)
Insert into #Product_list values
('banana'),
('strawberry'),
('orange')
SELECT * FROM REWARDS
join #Product_list
on ProductCsv like '%'+products+'%'

Answer 4

You could use a split function (there are many around on the web) to split ProductsCsv into individual elements and then compare that against banana, strawberry etc.
Ideally, you wouldn't store CSV data in a column, but instead have a separate table for that.