Prevent duplicates for indexed attributes

Currently, we rely on the caller to eliminate duplicate values. This works fine. But it'd be nice to prevent that at the database level just to be safe.

This adds a _small_ amount of overhead, but it's relatively trivial compared to the benefit. With this `OUTER APPLY` added, attribute values will _only_ be added if they are different from the last value saved for that key. This ensures we don't introduce unnecessary consecutive duplicates to the attributes table.

This is handy for casual scripting, as we can call `UpdateTopic` with whatever values we want, and ensure they are added if appropriate. If they are already the current value, they will be skipped. This also adds utility for cases where we are importing data, but don't know the current state of the database, as it prevents the need to first query the database, then deduplicate the records, and finally the delta.

Author: JeremyCaney

OnTopic.Data.Sql.Database/Stored Procedures/UpdateTopic.sql

@@ -32,9 +32,18 @@ SELECT	@TopicID,
 	AttributeKey,
 	AttributeValue,
 	@Version
-FROM	@Attributes
+FROM	@Attributes		New
+OUTER APPLY (
+    SELECT	TOP 1
+	AttributeValue		AS ExistingValue
+    FROM	Attributes
+    WHERE	TopicID		= @TopicID
+      AND	AttributeKey		= New.AttributeKey
+    ORDER BY	Version		DESC
+  )			Existing
 WHERE	AttributeKey		!= 'ParentId'
-  AND	IsNull(AttributeValue, '')	!= ''
+  AND	ISNULL(AttributeValue, '')	!= ''
+  AND 	ISNULL(ExistingValue, '')	!= AttributeValue
 
 --------------------------------------------------------------------------------------------------------------------------------
 -- PULL PREVIOUS EXTENDED ATTRIBUTES