If a target index does not build, a message about duplicate keys may appear. This may be due to case sensitivity of the source and target servers. For example, an RMS server may read "JOHN DOE" and "John Doe" and "john doe" differently, but a SQL Server may read them equally.

Although a unique index may exist on a source database, indicating that the data is unique, the server may not accept the data and declare that it is not unique.

To analyze this problem, first perform a query consisting of concatenating the key columns together and checking for counts greater than one. Such a query might show 4 duplicate rows for a table with unique key columns Name and Address, for example. To solve the problem, correct the data on the source server. If (for instance) "John Doe", "JOHN DOE" and "john doe" are really the same person, merge the information in the three records together.

In cases where part number "aaa-2756" is, in fact, distinct from part number "AAA-2756", build SQL Server (or whatever other target server we are using) in case-sensitive mode.

If an incremental synchronization gives an error about duplicate keys, the error message provides a sample query that should return the duplicate data to aid in resolving the issue. Insert the query directly into the InfoNaut query tool to locate the duplicate data.