Is your feature request related to a problem? Please describe.

I am trying to drop particular indices from a dimension that doesn't have coordinates.

Following: [drop_sel() documentation](http://xarray.pydata.org/en/stable/generated/xarray.Dataset.drop_sel.html#xarray.Dataset.drop_se

Some regular grids in Iris are generated from a reference point along a given coordinate direction along with an associated spacing.

The way this is currently implemented uses numpy.arange as follows:

https://github.com/SciTools/iris/blob/b2ce2a34e2eef7e3d6203c77ada7ed4ce89e3145/lib/iris/coords.py#L2293

Where zeroth is the value prior to the first point value, step is the numer

I am trying to learn to use siphon. It seems very powerful. Thank you for creating it!

My usual method for learning a new library is to try to introspect the objects via a combination of doc strings, tab completion, etc (in addition to rtfm of course). One challenge with siphon is that it's __repr__ methods for its objects are often insufficiently informative, making it hard to quickly unders

Does CLUST_WTS not exist in GDL or am I missing something?

GDL> array=dist(500)
GDL> weights = CLUST_WTS(array, N_CLUSTERS = 500, n_iter=10)
% Function not found: CLUST_WTS

Random number functions work best when they generate many numbers are once, which isn't always convenient. Python generators offer a solution to this, but they aren't conveniently carried around. Create two random number generator functions, one normally distributed, the other in a top hat distribution, and attach them to the pysat object at instantiation.

Plan:
- Create random number gene