There's no way logically that can work because you would be joining all three columns on all three columns. The syntax doesn't support it and the language doesn't either.

Can you...

You would have to poll all the tables in the database, one-by-one, to determine if there are changes. That's a poor approach. You could read the log using DBCC LOG...

Going to a heap in the hopes that you get improved compression is not the way to tune queries. If that clustered index is in use, it needs to stay...

A view doesn't have statistics.

A materialized view does, because a materialized view is an index. You can't drop statistics from an index.

I think I'd add another table in there. Instead of having a direct link between ProductPropertyType and ProductPropertyValue, I would define the values, the properties, and then have a table...

Eric M Russell (9/27/2016)

---

Talib123 (9/27/2016)

---

Hi All,

I'm a production DBA and have been for a good while.

I can query, join filters but how do I get my T-sql skills to...

There has to be something else causing the issue. If you're seeing different execution plans, it's because the mechanisms used to generate those plans have different inputs.

psred (9/27/2016)

---

The DB is exact copy of source(Query runs without filling tempdb) that gets copied daily morning to destination(Query fills tempdb). So stats, DB settings are same. Coming to server...

psred (9/27/2016)

---

The DB is exact copy of source(Query runs without filling tempdb) that gets copied daily morning to destination(Query fills tempdb). So stats, DB settings are same. Coming to server...

Then logically, you have an...

If you're getting wildly different plans between servers, there has to be differences. The differences can be in the statistics on the data in the database. The differences can be...

Rich Mechaber (9/27/2016)

---

But. If I worked for Grant, I'd be terrified of telling him anything in the way of bad news. Who's gonna volunteer to inform "The Scary...

Yeah, Columnstore is a good thing is you're doing lots of aggregates in the reports. If we're doing point lookups & just general data retrieval though, it falls down badly....

Also, what is the Recovery Point Objective, the amount of data that the company is willing to lose during the log shipping operation. The answer is always zero, but zero...

The same general rules are going to apply to this as they would to any other set of indexes. You're talking about having 1-16 different criteria for searching the table....