I've implemented DbDataAdapter to support filling when the schema already exists.
According to MS implementation, if the column already exists, it is unmodified (even if datatype is wrong). However, complications come when there are multiple columns with the same name (I don't know whether it would happen at all). So this new implementation stores a new ArrayList of source column names. Whenever a source column is scanned, calculate its correct name, adding a trailing number if required, and add it to this ArrayList. Then find the DataColumnMapping (do we need to calculate the DataTableMapping per row?!) to map it to the corresponding dataset column name. Then find this name in the datatable to see if we really need to add this new column. Tim, can you look at it to see if it is implemented correctly. Regards, Alan
DbDataAdapter.patch
Description: Binary data
