New session added! A peer-reviewed system for merging duplicates and their related records.Back to stream
Frank van der Most 1 week ago
Many databases have the phenomenon of doubles : two or more records in the same table representing the same data, whereas there should be only one. Dealing with such doubles is not always necessary, but when necessary not always trivial.
Preventing doubles from occurring is one way to tackle the problem but not always possible or perhaps not water tight.
For a research database containing data on Swedish innovations, I designed a system with which users can propose the merging of two records, compare them side by side, discuss the proposal, reach a conclusion and proceed with either merging the doubles or registering that the two records are indeed about two separate items ('false doubles'). The system takes care of data in related tables and keeps the administration of the accepted and rejected merge proposals for later reporting and preventing re-reviewing of false doubles.
This is my first attendence to the .fmp conference, so all feedback in advance is most welcome. If you're interested in particular details then let me know here, so I can try to include them in the presentation.