Most of my post-processing is going to be done with Jupyter/Pandas which handles Date-Time formats very nicely. If using external data, these timestamps should be used to define the maximum time-step. Smaller time-steps could be performed by only reported to output upon user request. For all other cases where the loads are being generated from the profile objects, the time should be referenced from some start time, like 2000-01-01 00:00:00, or similar. Regardless, the output processor should index by the timestamp and not just a simple index.
Most of my post-processing is going to be done with Jupyter/Pandas which handles Date-Time formats very nicely. If using external data, these timestamps should be used to define the maximum time-step. Smaller time-steps could be performed by only reported to output upon user request. For all other cases where the loads are being generated from the profile objects, the time should be referenced from some start time, like 2000-01-01 00:00:00, or similar. Regardless, the output processor should index by the timestamp and not just a simple index.