fiberIds are incorrect in (at least) mcsData and CobraConfig tables

Looking at the data for e.g. visits BETWEEN 1762300 AND 1762349 I discovered that the fiberIds in the database are incorrect.  For example, for fiberIds 102..103 the figure shows that the spots at e.g. y = 960 are assigned to fibres 102 and 103.

The problem also shows up in the mapping of nominal positions to fiberId:

Note that as a function of visit number the fiberIds change