Hw 4 #24, 11 about comp-econ-sp19 HOT 2 CLOSED

Since you select out the year 2000 (df.YEAR == 2000), you don't then need to group by year. You can, and that's why it still runs ok at that point. The problem arises when you try to use weights_2000 as the weights. When you group by year, it puts YEAR in as a level in the MultiIndex. Just don't group by year, and that should help!

from comp-econ-sp19.

Actually I had the same bug, but I think it is because you should look at the "YEAR" variable more closely. "YEAR" is actually a date... not a 4 digit year which tripped me up. I think we should set df.YEAR == '2000-01-01' instead of 2000.

Hope this makes sense and fixes errors

from comp-econ-sp19.