results - Results.read_scenario_data reads in incorrect data

[willu47]

results.read_scenario_data reads in the wrong data, and the data changes as the list of timesteps passed changes. Somewhere, there is a misalignment between the Spec coordinates and the data.

For example, actual data:

timestep	lad_uk_2016	residential_or_non	floor_area
2030	E06000056	residential	10841750.0
2030	E06000056	non_residential	6525.1228990099

Using the api

>>> df = results.read_scenario_data(
        'socio-economic', 'arc_baseline', 'floor_area', [2030])
>>> df[df['lad_uk_2016'] == 'E06000056']

returns the incorrected data:

timestep	lad_uk_2016	residential_or_non	floor_area
2030	E06000056	residential	3424711.285
2030	E06000056	non_residential	4647666.550

Direct read from the store gives expected data:

>>> store = results._store
>>> df = store.read_scenario_variant_data('socio-economic', 'arc_baseline', 'floor_area', 2030).as_df().reset_index()
>>> df[df['lad_uk_2016'] == 'E06000056']

lad_uk_2016	residential_or_non	floor_area
E06000056	residential	1.084175e+07
E06000056	non_residential	6.525123e+03

[fcooper8472]

Associated changes also needed in the results API, which is now more thoroughly tested.