An important use case that keeps coming up for large datasets is that they don't fit in memory, so it is convenient to run GLS on partitions of the data separately and save those results on disk. It is not obvious to end users how to then perform the cross-partition analysis from those individual GLS components. Some new functions should be introduced to handle this.
An important use case that keeps coming up for large datasets is that they don't fit in memory, so it is convenient to run GLS on partitions of the data separately and save those results on disk. It is not obvious to end users how to then perform the cross-partition analysis from those individual GLS components. Some new functions should be introduced to handle this.