Erik Darling is answering questions again:
My company (using SQL Server 2019 Standard) has an algorithm that keys addresses into a varchar(40) has a cross-reference database that assigns an identity property to each new value, allowing us to post the numeric in our datasets. Production has to search their generated string keys in this database’s table to get the integer key in return. We have ensured proper string data typing on lookups and have unique compressed indexes on the string values. What would your next tuning step be if this was not getting the performance you needed?
There’s a good set of questions this time, so click through for Erik’s answers.