How come I can filter on a column that I have already dropped in spark?

Question

When I drop a column from a dataframe in spark, I expect that I can no longer use that column.

However, it turns out that I can in this example:

schema