:unsure: What cross join?

No real problem - the row count just isn't guaranteed to be absolutely spot on.

Oh and sysindexes is deprecated.

Sadly...no

Ha ha! Awesome!

Here's the slightly more optimised version I promised.

It uses an integer hash of the customer_full_name and postcode to quickly find potential matches without huge string matching.

Only potential matches are then...

Gopi,

That's good, but I don't think we can safely hard-code the IdCol values like that. I'm pretty sure there are more than 5 records...

Paul

Might be a similar explanation to this Connect Item:

https://connect.microsoft.com/SQLServer/feedback/details/437859/test-performed-on-filtered-out-rows

Optimiser behaviour changes between versions, so it could be that. Without a reproduction script that works for me, there's a limit...

Ok, I need to do some stuff with a hash and an index to make this fly for larger sets, but as a demonstration only:

Setup script (slightly modified)

CREATE TABLE...

Copy & paste, ran the script, and...

(0 row(s) affected)

No error message. That is on SQL Server 2008 SP1 CU8 build 10.00.2775

Hey Matt,

So that's reasonably clear, but what output do you want? Do you want a list of matches between the two tables on email address AND postcode, where there...

You generally want to avoid tables with such (potentially) large rows. Consider defining sensibly-sized variable-length columns, rather than just picking a huge number.

The warning is 2000-specific: this won't cause...

DECLARE @mytable

TABLE (

id INTEGER NOT NULL,

potype...

One more:

SELECT Derived.table_name,

row_count = SUM(P.[rows])

FROM sys.tables T

JOIN sys.partitions P

...

I have a feeling you tried quite hard there to make your point clear, but I'm here to tell you that you were not 100% successful 😀

Is it possible to...