Introduce new DeleteConsecutiveExtendedAttributes maintenance procedure

In older versions of the OnTopic Library, new extended attribute versions were _always_ written, even if their composite XML value had not changed. This isn't necessary in terms of the OnTopic library, which will take whatever the latest version of an attribute is when loading the topic graph (or restoring a topic). As such, these _consecutive duplicates_ only add clutter to the database, and potentially slow down some queries by introducing more data to sort through. This is especially concerning for the extended attributes since they use large XML fields. The `DeleteConsecutiveExtendedAttributes` identifies and deletes these from the database, thus making it more efficient. This corresponds to #fc3e27a, which does the same for indexed attributes.

Author: JeremyCaney

OnTopic.Data.Sql.Database/Maintenance/DeleteConsecutiveExtendedAttributes.sql

@@ -0,0 +1,71 @@
+--------------------------------------------------------------------------------------------------------------------------------
+-- DELETE CONSECUTIVE ATTRIBUTES
+--------------------------------------------------------------------------------------------------------------------------------
+-- Current versions of the OnTopic Library evaluate whether or not the composite XML for the extended attribute values has
+-- changed since the previous version, and only creates a new version if it has. This wasn't true in previous versions, however.
+-- As a result, there are some cases, and especially in older databases, where unnecessary duplicates occur for attribute
+-- values. These dramatically increase the size of the database and can slow down the processing time of certain queries. This
+-- procedure will detect concurrent duplicates and remove them from the database. This reduces the size of the database, without
+-- interfering with the data integrity.
+--------------------------------------------------------------------------------------------------------------------------------
+-- NOTE: Because this query must cast the XML values as VARCHAR in order to compare them, it takes a LONG time to run. Please
+-- be patient!
+--------------------------------------------------------------------------------------------------------------------------------
+
+CREATE PROCEDURE [dbo].[DeleteConsecutiveExtendedAttributes]
+AS
+
+SET NOCOUNT ON;
+
+--------------------------------------------------------------------------------------------------------------------------------
+-- CHECK INITIAL VALUES
+--------------------------------------------------------------------------------------------------------------------------------
+DECLARE	@Count	INT
+
+SELECT	@Count	= Count(TopicID)
+FROM	ExtendedAttributes
+
+Print('Initial Count: ' + CAST(@Count AS VARCHAR) + ' Extended Attributes in the database.');
+
+--------------------------------------------------------------------------------------------------------------------------------
+-- IDENTIFY GROUPS OF CONCURRENT DUPLICATES
+--------------------------------------------------------------------------------------------------------------------------------
+WITH GroupedValues AS (
+  SELECT	TopicID,
+	AttributesXml,
+	DateModified,
+	Version,
+	ValueGroup = ROW_NUMBER() OVER(PARTITION BY TopicID ORDER BY TopicID, Version)
+	- ROW_NUMBER() OVER(PARTITION BY TopicID, CAST(AttributesXml AS NVARCHAR(MAX)) ORDER BY TopicID, Version)
+   FROM	ExtendedAttributes
+),
+
+--------------------------------------------------------------------------------------------------------------------------------
+-- RANK DUPLICATES BY DATE
+--------------------------------------------------------------------------------------------------------------------------------
+RankedValues AS (
+  SELECT	TopicID,
+	AttributesXml,
+	DateModified,
+	Version,
+	ValueGroup,
+	ValueRank = ROW_NUMBER() OVER(PARTITION BY ValueGroup, TopicID, CAST(AttributesXml AS NVARCHAR(MAX)) ORDER BY TopicID, Version)
+  FROM	GroupedValues
+)
+
+--------------------------------------------------------------------------------------------------------------------------------
+-- DELETE NEWER DUPLICATES
+--------------------------------------------------------------------------------------------------------------------------------
+DELETE
+FROM	RankedValues
+WHERE	ValueRank > 1;
+
+PRINT('Concurrent duplicates have been deleted.')
+
+--------------------------------------------------------------------------------------------------------------------------------
+-- CHECK FINAL VALUES
+--------------------------------------------------------------------------------------------------------------------------------
+SELECT	@Count	= @Count - Count(TopicID)
+FROM	ExtendedAttributes
+
+Print('Final Count: ' + CAST(@Count AS VARCHAR) + ' duplicate Extended Attributes were identified and deleted.')

OnTopic.Data.Sql.Database/OnTopic.Data.Sql.Database.sqlproj

@@ -100,6 +100,7 @@
     <Build Include="Types\AttributeValues.sql" />
     <Build Include="Types\TopicList.sql" />
     <Build Include="Views\AttributeIndex.sql" />
+    <Build Include="Maintenance\DeleteConsecutiveExtendedAttributes.sql" />
     <Build Include="Maintenance\DeleteConsecutiveAttributes.sql" />
   </ItemGroup>
   <ItemGroup>