As specified in RFC-517, the schema for the DiaForcedSource table in the PPDB does not include the ra/dec at which force photometry was performed; that information is instead stored in the DiaObject record that was valid at the time of force photometry. It thus requires a semi-complicated join to extract the force photometry coordinates for a light curve, which could be important for diagnosing photometric issues if the centroid is moving substantially between epochs. 

At the time of RFC-517, we decided that this join is feasible enough, and the storage costs of adding ra/dec to DiaForcedSource are significant enough, that this schema is acceptable. However, as our prototypes of the PPDB become more advanced, we may wish to revisit this question with more quantitative performance measurements.

This ticket is meant to remind us to reevaluate this topic.