bug: incorrect saturation when querying budget file with full3D=True

[aestrad-intera]

The following code returns an array of all zero's:
cbb = flopy.utils.CellBudgetFile(test_ws/'model.cbb')
sat = cbb.get_data(text="SAT", kstpkper=(0, sp-1),
full3D=True)[0].data[1]

however, If I run:
sat = cbb.get_data(text="SAT", kstpkper=(0, sp-1))

it returns a record array with node, node2, q, and sat. I have no idea what q is, but it is all zeros. Sat has correct saturation values. Please fix the full3D version to pull the actual cell saturations not just zeros.

[wpbonelli]

This is less bug, more awkward API. The IMETH 6 budget format always has a q column, which is meaningless when you're looking at aux vars like saturation. By default with full3D you get q. To get a 3D saturation array before now you'd need to not use full3D and manually reshape the recarray saturation column.

#2776 adds a variable param to get_data() like the one on get_ts(). Here it will only work with full3D. It's a bit awkward to have to say text="DATA-SAT" and variable="sat", but I'm hesitant to infer column based on text and the value of the "q" array; that seems brittle.