Improve loading with "name" string to return all available data

[abooton]

As a user of netcdf-CF compliant data, I want my data to always load when I provide a valid "name string" in the load command, so that I can be confident it will be loaded regardless of the naming attributes it has.

Acc Criteria:

• When supplying a "name string" on load, i.e. iris.load(filename, name_string) all matching data should be returned regardless of whether name_string matches the standard_name, long_name or var_name. (i.e. if the data has a standard_name and a long_name it should be loaded even if the user supplies the long_name).
• Include an example of loading a cube using the "name string" (as in the above case)
• Include an example of how to load a cube in a manner that replicates the current behaviour, so users can easily retain our original functionality should they needed to.