Fix multi-output features not created when there is no child data

[jeffzi]

Fix multi-ouput features not created when there is no child data

When there is no child data, calculate_feature_matrix raises a KeyError because multi-output features are not created. The expected behaviour is to have those features represented by columns filled by numpy.nan, as it is the case with regular features.

Here is minimal reproducible example:

import numpy as np
import pandas as pd

import featuretools as ft
from featuretools.primitives import NMostCommon

parent_df = pd.DataFrame({"id": [1]})
child_df = pd.DataFrame({"id": [1, 2, 3],
                         "parent_id": [1, 1, 1],
                         "time_index": pd.date_range(start='1/1/2018', periods=3),
                         "cat": ['a', 'a', 'b']})

es = ft.EntitySet(id="blah")
es.entity_from_dataframe(entity_id="parent", dataframe=parent_df, index="id")
es.entity_from_dataframe(entity_id="child", dataframe=child_df, index="id", time_index="time_index")
es.add_relationship(ft.Relationship(es["parent"]["id"], es["child"]["parent_id"]))

n_most_common = ft.Feature(es["child"]['cat'], parent_entity=es["parent"], primitive=NMostCommon)

# cutoff time before all rows
# We expect N_MOST_COMMON features to be np.nan
ft.calculate_feature_matrix(entityset=es, 
                            features=[n_most_common],
                            cutoff_time=pd.Timestamp("12/31/2017"))
#> [ ... ]
#> KeyError: "None of [Index(['N_MOST_COMMON(child.cat)[0]', 'N_MOST_COMMON(child.cat)[1]',\n       'N_MOST_COMMON(child.cat)[2]'],\n      dtype='object')] are in the [columns]"

^{Created on 2019-12-05 by the reprexpy package}

This PR fixes the bug and adds a test for multi-output feature in test_empty_child_dataframe

[rwedge]

the '#' character should be removed

[jeffzi]

My bad, I fixed it.

[rwedge]

I think the PR looks good, once the PR fixing the issue with sklearn and the tests goes through I think this will be good to go.

[codecov-io]

Codecov Report

Merging #834 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #834      +/-   ##
==========================================
+ Coverage   98.15%   98.16%   +<.01%     
==========================================
  Files         117      117              
  Lines       10848    10851       +3     
==========================================
+ Hits        10648    10652       +4     
+ Misses        200      199       -1

Impacted Files	Coverage Δ
...mputational_backend/test_feature_set_calculator.py	100% <100%> (ø)	⬆️
...s/computational_backends/feature_set_calculator.py	98.55% <100%> (ø)	⬆️
...computational_backends/calculate_feature_matrix.py	98.58% <0%> (+0.35%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cb82feb...849778a. Read the comment docs.