we had to drag info out of the original poster and Grant just googled "in memory open source" - ... i think he might have given up on...

Little bit of amusement for you all. Was going through some old processes, that out date me at the business, they're being "questioned" (aka, people have forgotten what they...

looks like this is Grant's winning argument to Extended events. , but I will advise that you need to know your hardware setup before you measure disk I/O and whether...

I'm doing to be honest here, it looks like you are looking for a solution before you know what the problem actually is... we can't help you without details

so you want open source - which means "free", beware of the licence- you may have to disclose your code which may be not in the interests of your company

as...

I would advise against running in high safety mode if possible (if your mirror server falls down then you can't commit transaction on the principal) - I've always run in...

since I couldn't see the bottom of the plan, very stupid question - are there any triggers on the table?

Occasionally in this scenario I've used a computed column (normally I persist it - PCC) that mimics your lookup values. once you get all of your code sorted (joins, views...

So, please, I hope, you're testing this in a non-production environment, because, here's a bit of fun trivia, once you mark a database as in-memory, you can never undo...

as I mentioned memory optimised tables need to be treated with care.

what are you trying to fix? have you got stats that suggest your issues?

I think I found the issue, not sure why , but the "instead of" trigger on that table seems to be re-firing itself

I turned the "allow triggers to fire triggers"...

I managed to track down the code in the trigger (using print statements on my 2019 dev kit)

INSERT INTO order_line
(
    account_code,
   ...

yes,

there is also a begin tran/try catch/commit/rollback in the calling proc

I know the proc should be handling the transaction, but someone has been a bit "slap happy" with XACT_ABORT and...

Ultimately memory optimised tables end up having to be persisted to disk and lord alone I've never tested them with bulk inserts and data manipulation

Are you maybe talking about using...

apologies - I read "truncate table" as "truncate log" my mistake