you don't have any filtering on #SOURCETABLE - you only filter on the join, so all records from #SOURCETABLE  will be selected.

and the only record you have on the sample...

according to documentation it would seem that it does not but worth confirming with someone that does work with PDW

https://docs.microsoft.com/en-us/sql/relational-databases/system-stored-procedures/database-mail-stored-procedures-transact-sql?view=sql-server-ver15

sysadmin is server (instance) level role as seen on

description of what each column for syslogins below

https://docs.microsoft.com/en-us/sql/relational-databases/system-compatibility-views/sys-syslogins-transact-sql?view=sql-server-ver15

and as you will need it for sure https://docs.microsoft.com/en-us/dotnet/framework/data/adonet/sql/server-and-database-roles-in-sql-server

Ideally you would load the events onto a sql server table and do the analysis and initial filtering there - easier with high volumes.

sample example that I use to do...

then don't commit the username/password - that could be on a "do not commit" file on your machine and get BIML to retrieve the username/password part from that file

in dev it should not matter - in all other environments all connections should be either done through the SSIS Catalog as sensitive or stored on a secure configuration database,...

duplicated post - reply in https://www.sqlservercentral.com/forums/topic/index-23

if you have a single filegroup with data files then you can't really tell it - some pages may be on one file, other pages on another file. So although...

• This reply was modified 5 years, 9 months ago by frederico_fonseca.

what everyone else said - plus in some of the teams we have a database per change request being done.

one of our teams has almost always over...

• This reply was modified 5 years, 9 months ago by frederico_fonseca.

Yes, there will be some limited page splits.  But the reads against the table will be vastly more efficient, even with the splits.

It's certainly possible that [TimeDayId] then [Material_Werksdaten_key]...

You should cluster the table on:

( Material_Werksdaten_key, TimeDayId )

Your join is very confusing.  You need to specify the TimeDayId(s) directly, not use a function on it.

Is TimeDayId one value...

ah well. didn't notice SQL Saturday was this near me (I'm in Dublin) - would have gone there if nothing else to talk with some people from here.

what everyone else said - plus in some of the teams we have a database per change request being done.

one of our teams has almost always over 100 active databases...

Based on what you said I would change the table itself and have 2 new columns made up of the year and and month part of TimeDayId so the index...

thats SSIS for you - happens a lot unfortunately and I never managed to find a solution.

try with the 2016 runtime and see if it improves. https://www.microsoft.com/en-us/download/details.aspx?id=50040

if it still fails...