tests are failing due to test_decimal_with_joins

[fdosani]

Seems like the TravisCI tests are failing. I think this is due to the version of various packages not being pinned. As the packages change and evolve so will testing behaviour.

The following test is failing:

def test_decimal_with_joins():
    df1 = pd.DataFrame([{"a": Decimal("1"), "b": 2}, {"a": Decimal("2"), "b": 2}])
    df2 = pd.DataFrame([{"a": 1, "b": 2}, {"a": 2, "b": 2}])
    with raises(ValueError):
        compare = datacompy.Compare(df1, df2, "a")

I think in the past Pandas might have has different behaviour. Testing with 0.19.0 version of Pandas and Python 3.5 it still seems to NOT throw any error.

[fdosani]

Looking back in the issue history we did talk about this, and I guess it was reverted back to pre-0.23.x functionality in 0.24+. Ref: #19

I think removing the raises is the right thing to do. Seems to be a pandas versioning issue/bug.
Pandas Ref: pandas-dev/pandas#21681