Best question ever: dba.stackexchange.com/q/324910/150011

@Charlieface I don't believe you can express it directly. Aside from relying on the output order of a FULL MERGE JOIN, you might be able to express it as dbfiddle.uk/_ddCBZLv

Assuming your main concern is avoiding a sort while guaranteeing order

Anyone running SSMS 19 seriously? I keep seeing issues with it, so haven't tried it yet

I’ve been using it forever

@ErikDarling Alongside 18 or instead of?

A chairde - Morning all!

@ypercubeᵀᴹ - don't know if you noticed this - will report on Monday!

@Vérace See postgresql.org/docs/current/typeconv-union-case.html

OMG - I'll have to investigate further - you've actually read the heretical texts of the PostgreSQL Ayatollahs! Fair play for diversity of knowledge - you'll probably be condemned to a life of MySQL db maintenance for your pains! :-)

Well, all database engines have to solve or address many of the same fundamental issues

@PaulWhite Can't do replication via the UI in it, so I'm running 18 still too. 🙃

I think they've addressed most of the replication issues in minor updates but there might be one or two still outstanding yeah

Yeah... and some of them don't really address them!

@PaulWhite Thanks for not blaming replication on this one. 😉

I think it's implied

@Vérace I did notice the comment. Not enough time for anything further.

what the heck is that word lol

@PaulWhite instead of. completely uninstalled 18 when 19 dropped. no issues to report, and i'd imagine we have pretty similar usage patterns.

unless you're secretly working with replication and powerbi

🔳🔳🔳🔳🔳

@ErikDarling what about SSIS and FileTables too? 🫠

@PaulWhite Something like that was tumbling around in my head but couldn't put my finger on it. Annoying you can't specify it explicitly. Logically the compiler should be able to infer ISNULL(id1, id2) to mean the same thing.

Yeah

Yeah