That's more or less what I did in January on a 1.2 billion record table, that had no other tables related on the identity column.  I actually used SSMS data...

Yes I had to do that on the first table in the list last January, but in that case the identity integer column being upsized didn't exist anywhere else.  For...

Run times during the weekday business hours can vary from a few seconds to 2 minutes generally.  Logical reads actually seem small compared to many of our ORM-generated queries.  But...

Thanks, that's just the "ammunition" I needed.   You South Africans have an influence way beyond what the size of your country might suggest   🙂    I'm also a...

Well our whole production sql server system is on Netapp Enterprise solid state drives now, and of course QA isn't.  Actually we do have one QA environment quite similar to...

In my case real life would be a buffer cache stuffed with over 400GB of data ( data my query needs? who knows? )   Also real life would be...

Isn't your cache essentially cold whenever a given query has to do physical reads? Unless you can hold the entire DB in memory....

I read Jonathan's article.  Disappointing that such a basic counter is suspect.   And even though 470GB of memory sounds like a lot, it's only 6% or so of the...

Ok that's what I thought.   So with that being near 100% ( except sometimes during evening index maintenance ), it's not clear how much help we're going to get...

Here's the latch timeout info.   I'm attaching an image showing memory use from our Spotlight tool.  The graphs have legends but for the one called "Cache" the explanation is:

I started this index analysis project after spending a number of months removing unused indexes and applying the highest value missing indexes ( mostly as identified by sp_blitzIndex when the...

If you comment out the RaisError portion and try to use THROW instead, it stops after one loop.  I'd like to use THROW and have it continue the while  loop.

Yes with the new index we get a seek and runs in 50 seconds, then once cached, a few seconds.  I think part of what is happening is the clustered...

ok thanks.   I'll try disabling the single-column index and adding one sql is suggesting for this query.  Since this is a large table with many columns ( over 100...

I updated stats with fullscan on the two indexes being ignored.  Still runs for six minutes and ignores the multi-column index which, when forced with a hint, cuts the runtime...