I'm trying to do a column by column, row by row comparison for 75 tables in two different databases. The tables contain up to 165 million records.
I need to determine if there are any missing records, and we need to identify any columns that contain different values. Potential problems include decimal errors.
I would start with rowcounts. At least it gives you an idea of missing records. I know the counts could match but the data would be different, but it's a start. Maybe create some sums on columns to verify