My understanding is that in GR the way in which the components of the metric tensor can vary from location to location in spacetime is determined or constrained by a number of factors: (1) the choice of coordinates (a mathematical constraint); (2) the distribution of mass/energy (a physical constraint that is the heart of GR); and (3) the requirement that special relativity rule locally (a logical constraint also at the heart of GR). In specific situations there are, fourthly, additional constraints required to match observations (observational constraints). And there must be other constraints I am unaware of; e.g. perhaps the components of the metric tensor must be single-valued functions of whatever coordinates are chosen? I find it difficult to appreciate just which features of a metric used as a tool for a particular purpose --- of the finished product, as it were --- are associated with what constraint. I need help in the case of the Robertson-Walker (RW) metric of cosmology used for the accepted LCDM model. My understanding is that this metric is chosen subject to the additional constraints of: (5) the Copernican Principle, to match to a postulated homogeneous, isotropic universe; (6) a need to be able to define a universal time coordinate (a mathematical constraint?); (7) a need to account for the redshift (an observational constraint); (8) the simplifications introduced by the now-accepted Euclidean character of space sections (a recent and welcome constraint?); and finally, (9) the assumed and observed ubiquitous rule of the known laws of physics throughout spacetime, together with their incorporated constants, in particular c. I would like to better understand just which features of the RW metric are uniquely and fully determined by the constraints I have listed above ... and by others I may have missed. In particular, is the choice of a scale factor to describe change uniquely mandated by these constraints? I suspect not, but would like to know for sure.