That's a good question. While DB Checkpoints in TestComplete are easy to use and convenient, because information is so variable in many of my testing projects, I prefer to write my own code for comparing database values. Running an SQL query against a table, outputting it either into an in memory object or into some sort of file (CSV is what I've used in the past) and then comparing the output to a baseline value somewhere has been my method. It's what I used before DB Checkpoints were a thing and has worked very well. It resolves your problem with having to alter the DB source.
One of the things I have in my "to-do" list for my SQLUtilities script extension is to add some sort of checkpoint functionality to allow a greater exposure to the SQL query and the data source. I hope to have the time to get that done by the end of April. In the meantime, take a look at using the ADO objects available in TestComplete for building SQL queries and doing your own data validation.