Partitioned datasets need descriptive labels in the pimc.h5 file.

[IanGalbraith]

As suggested in the email.

If you take the average of the first energy dataset divided by the first weight dataset, that should give you energy.