Possible causes: Approximate numeric datatypes include float, double precision, and real. The exact value of an approximate numeric datatype can vary from one platform to another, and can cause comparison errors such as:
- If a key column includes an approximate numeric datatype, there is no guarantee of the DA server matching the source and target columns. Each failure to do so creates two false differences: one “missing” row in the source database and one “orphaned” row in the target database.
- If a set of columns for comparison include an approximate numeric datatype, there is no guarantee the two matching source and target rows are considered to be equal. Each failure creates a false “inconsistent” difference.
Solution: None.
You may be able to avoid false differences by skipping approximate numeric datatypes when creating column mappings—although this introduces the risk of genuine differences between two approximate numeric datatype columns that may go unnoticed.
Note: Reconciliation cannot fix false differences.