Run-time creation of Repeating Stream Sets
Repeating stream sets have made an important contribution to csv2json functionality and performance. In particular, the repeating streamset concept has eliminated the use of bitwise logic to compile the fixed field name template into the csv2json pipeline. As a consequence, the current csv2json has dramatically improved performance for csv files with very large numbers of columns.
Further progress in eliminating JIT compile-time dependencies on CSV file type (following the general concern of issue #35) can be achieved by a capability for creating repeating stream sets at run-time rather than JIT compile-time.
The simplest approach is to create a method that converts a fixed-length input stream set into a repeating stream set of the same shape.