You could use the following terminology, which would be readily understood by anyone who works with structured data in the common interactive tools.
Given a collection of elements that can be unravelled or flattened into a list:
hard reshape assigns the elements of the list to a new structured variable (essentially a list of lists) where the lengths of the component lists are determined by a “shape vector” passed as an argument, and if the number of elements is insufficient for the construction, the assignment fails.
soft reshape, where the assignment does not fail, and where the “missing data” is filled with padding elements, repetition of the original list elements, etc.
The main problem with this approach is that the rules for soft reshape are often hard to nail down exactly. Some users want the rules to work one way, and others want them to work some other way.
You might be better off sticking to a single reshape function implemented as a hard reshape, but allowing the shape vector to be calculated by a function.
In my view, difficulties in naming frequently reveal a deeper problem, as much in software as they do in ordinary English writing.