Multi-target Query Assistance

Is there a certain structure to the breakdown of data in Content1 and Content2, and what is the reasoning of them being two separate tables (they appear to have the same definition)?  For example, are newer records (by CreatedDateUtc) in Content2 and older records in Content1?  Also, what was the bottleneck with performance in your attempts so far?  What do the queries look like?  Understanding this context should help with solutioning.

Perhaps there is a reasonable partitioning scheme that would sufficiently optimize the lookup?  There are some potential drawbacks, but maybe an indexed view would be worth consideration in this scenario as well.

I should have prefaced this with "This is a vendor database and we do not own the schema."  But that % would have my first question if I saw this question in the wild, itCanOnlyBeJared .

KMSigma.SWI said:

I should have prefaced this with "This is a vendor database and we do not own the schema."

Ah, fun!  While you don't own the schema, are you free to add indexes or views as long as there aren't any drastic changes that would impact the application?  Or are you limited to coming up with the most efficient lookup query possible with exactly what you've been handed?  With lookups over millions and millions of rows, the limited flexibility definitely hampers optimization attempts. 

We can add anything as long as we don't touch the original table content.  So yes, Indexes, views, functions, are all allowable.

I've tried a very ugly view (using the 28 connections with a LEFT JOIN), tried a multi-statement table function to be used as a CROSS APPLY, and a handful of other things.  None of them seem particularly performant.

How often are these tables updated?  Perhaps the indexed view option could still be worth a shot?  Are the join column(s) indexed?  Would it be possible to show the attempted view query and associated query plan for additional context?

How often are these tables updated?  Perhaps the indexed view option could still be worth a shot?  Are the join column(s) indexed?  Would it be possible to show the attempted view query and associated query plan for additional context?

That's a long story, but for the sake of argument let's say it's updated once a day.  Like I said, we can add to the schema, but can't touch the data or fields in the data.  But we can add indexes.

The indexed view option doesn't modify the data or fields of the underlying tables, but it does introduce schemabinding and prevents structural changes to the underlying tables that may impact the indexed view unless you drop the view first.  Some additional info with sample queries in this article:  https://www.sqlshack.com/sql-server-indexed-views/

The advantage is that the data in the view's index will be maintained for you, and if that is structured in the way you want to see your results then it should result in a simple and performant query plan.

There may be ways to optimize with indexes alone, but I'd have an easier time assessing further with example queries/indexes/plans.

A few clarifying questions...

1. Can you elaborate on the 28 possible connections between source and targets?
   
   1. You mentioned an ugly view using these with a LEFT join...maybe I'm missing some context regarding the complexity of these.
2. Are the necessary JOINs already indexed well?
3. If you just need to pull in the additional column values from the target tables where you have a matching key with the source table, have you tried two INNER JOIN queries (one for each target table) with a UNION between them (potentially inside a view definition)?  With indexes on the join columns, I'm curious if you'd get the performance you need with this approach.

Yes, the ContentID fields in each of the "connected" tables are indexed.

This is the framework I've done with that crazy join list...

SELECT [Contents].[ContentId]
     , COALESCE(
         [Content1].UserId
       , [Content2].UserId
       -- Additional fields here for the coalesce
       ) AS [UserId]
     , COALESCE(
         [Content1].CreatedDateUtc
       , [Content2].CreatedDateUtc
       -- Additional fields here for the coalesce
     ) AS [CreatedDateUtc]
FROM [ContentsSourceTable] AS [Contents] WITH (NOLOCK)
LEFT JOIN [custom_Content1] AS [Content1] WITH (NOLOCK)
     ON [Contents].ContentId = [Content1].ContentId
LEFT JOIN [custom_Content2] AS [Content2] WITH (NOLOCK)
     ON [Contents].ContentId = [Content2].ContentId
-- add a while bunch of other joins here for the 26 other types

I should note: I only play a DBA on television.  I'm a novice in many ways.

Thanks for the clarification and the sample query.  The LEFT JOINs in your query would be retrieving ALL values from the big table, and only matches from the smaller tables.  Since you are only looking for the ones with matching keys (eliminating NULLs from non-matches with COALESCE), I think the INNER JOIN approach with UNIONs might be a lot more performant.  Not sure if this is what you're looking for, but hopefully it is a step in the right direction.  Let me know if this helps.

SELECT [Contents].[ContentId]
     , [custom_Content1].UserId
     , [custom_Content1].CreatedDateUtc
FROM [ContentsSourceTable] AS [Contents] WITH (NOLOCK)
INNER JOIN [custom_Content1] WITH (NOLOCK)
     ON [Contents].ContentId = [custom_Content1].ContentId
     
UNION

SELECT [Contents].[ContentId]
     , [custom_Content2].UserId
     , [custom_Content2].CreatedDateUtc
FROM [ContentsSourceTable] AS [Contents] WITH (NOLOCK)
INNER JOIN [custom_Content2] WITH (NOLOCK)
     ON [Contents].ContentId = [custom_Content2].ContentId
-- add UNIONs here for the INNER JOIN queries to other 26 tables

Wanted to test this out when I had fresh eyes.  Today, I ran an experiment on the LEFT vs. INNER JOINs with my data set:

• Returning 1,000,000 rows with the LEFT JOIN: 12 seconds
• Returning 1,000,000 rows with the INNER JOIN: 0 seconds, but no rows returned.

Sadly, my initial hunch was correct, and I need to use the LEFT JOIN to eliminate bad lookups.

But thanks itCanOnlyBeJared for the recommendation.

And oddly, the speed seems way up this morning.  Might be the reindexing, might be the underlying IO, might be divine intervention, but I'll take whatever improvement I can.

When you tested the INNER JOINs, did you update the query to use the UNIONs as well (between the different selects for each target table)?  Substituting INNER for LEFT joins in the original query would have a much different effect.