I was recently troubleshooting a severe query performance issue when I ran into a query plan that used a nested loops join to combine data from two tables. The join was 97% of the cost. When I looked closer I saw that there were no join predicates specified. The join in the query looked pretty straightforward to me and used a single column on each side. Why would the optimizer do what equated to a cartesian join?
It turns out the query was a little more complicated than it looked. I was actually joining two views that were part of a view-based (non-persisted) dimension model. I was joining on logical primary and foreign keys which made sense.
When I dug into the views I found that one of the key columns was based on a user-defined fuction that was converting a datetime to an integer value so that it could be used to join to a date dimension view. Once I changed the underlying view to apply the same logic as the UDF at the view level, so that a call to the UDF was unecessary, the query executed within a second as expected.
Other behavior that I noticed was that if I changed the inner join to a left join, the optimizer came up with a different much more efficient plan. This appears to be a flaw in the optimizer but I would like to speak to someone at Microsoft before making that claim.
The lesson learned here is that if you have a poorly performing query due to a nested loops join with no join predicate(s) it's not ncessarily the case that you're missing join criteria in your query (as all the posts I was able to find seemed to point to). The culprit could be a UDF on a join column.
Awesome post. You saved me a ton of debugging time, and a query that was taking a half hour to return 17 rows got reduced to under a second just by moving the UDF code into the view.
ReplyDeleteI have also noticed in some cases changing an INNER JOIN to a LEFT JOIN results in an order of magnitude decrease in query plan cost. Something odd seems to be happening there.
ReplyDeleteGreat tips, thanks for sharing for the world to learn from. In my case, I had a view with a join clause that had a CASE statement in it. In one case it joined to one table, another case it joined to a different table. Imagine how the poor optimizer handled that.
ReplyDeleteThanks for the tip. I was trying to optimize a simple query which looks OK. When I noticed the warning 'Nested Loops Join - No Join Predicate' in the execution plan, I started digging and eventually landed here.
ReplyDeleteI could not get rid of the function (due to business logic), I created a Temp table and first saved the results of the sub query with the function. The optimizer now have a physical file to look at, and worked as expected.
OMG...thank you. I've spent an entire day on trying to figure out why I kept getting the Nested Loops - No Join Predicate....turns out I just had to change 2 of my joins to left joins. Luckily, it doesn't matter if I get nulls from these tables but this could be very aggravating if it did...
ReplyDeleteYou saved me a lot of time, thank you.
ReplyDeleteI had the same issue with a calculated join [join( select…) A on A.x=B.x].
select * from data
join (select …) A on A.DayDate=data.DayDate
Only the tipp with a tmp-Table worked, no CTE nor a dedicated view out of the inner join statement worked.
But changing the order of the joins and using a CTE like
with CTE(DayDate) as (select …)
select * from CTE
left join data on data.DayDate=CTE.DayDate
worked also