I do not have need for foreign language characters or collation since my data is received as pure ascii.
My concern is performance in the application code, and though I have read many arguments that varchar performs better than nvarchar, they did not discuss the overhead converting back and forth between Windows Unicode and SQL Server non-Unicode varchar data?
What is the bigger impact in general - performance within SQL Server dealing with varchar vs. nvarchar, or performance in the application, converting back and forth between Unicode on the Windows side and non-Unicode on the SQL Server side? I understand there are scenarios where one would perform better than other, so my main scenario is simple text search and retrieval through a couple million records. I am using Entity Framework 6.1.3 as the DAL.
This is a concern and I would love some feedback on this from anyone having experimented or knowledgeable on this topic. Articles discussing varchar vs. nvarchar act like the SQL database is totally isolated in respect to performance and external software does not come into play. Is there in fact Unicode impedance mismatch between Windows native use of Unicode and ascii in the database, storage space and transfer speeds ignored, that negates the ascii database performance benefits? Especially in relation to EF? Thanks