I am sorry but not much you can do without them fixing the extraction of the data.

Lets analyze the sample you gave us - which I trully...

Is that really the content of the file? if so it is a piece of crap data really.

ADV ,A_DAT ,A_TIM ,P_tit ,S_Tit
"SOCLEAN INC","5/16/2018","14:39:44",""Z-QUNICY,"DR K 60...

Take in consideration what Steve and Sue said and look at better/correct ways of doing your process.

However just as a possible rewrite of what you have you can...

on the counts replace the "else 0" with "else null"
e.g change
COUNT(distinct (CASE WHEN obranome LIKE 'C%' AND etotaldeb <> 0 THEN obrano ELSE 0 END)
to
COUNT(distinct (CASE...

If referring to PowerBI Report Server then any will do - but be careful with licensing - either you have a Enterprise edition with SA (and you can then install...

instead of presenting 8 vcpu with 1 core each present less vcpu with more cores. that will sort the licensing limit

you may be able to - but being a DMZ you probably should not anyway.

have a look at https://www.experts-exchange.com/questions/25053526/Passing-user-credentials-from-a-trusted-domain-through-SPN-and-Delegation.html. this contains other links that may help you...

did you install SSDT also?

https://docs.microsoft.com/en-us/sql/ssdt/download-sql-server-data-tools-ssdt

If your PowerBI is loading the whole table then following will most likely apply.
Even if not some of it may still be worthwhile doing.

PowerBI uses the...

any further char that you wish to include add them to the list

select *
from #TEMP t
where patindex('%[^A-Z ''-]%', name) = 0

[^A-Z ''-] means any...

select *
from #TEMP t
where patindex('%[^A-Z ''-]%', name) = 0

will do the trick here - but be aware that depending on your collation you may need more...