BUG: Identify SparseDataFrame as sparse

The is_sparse function checks to see if
an array-like is spare by checking to see
if it is an instance of ABCSparseArray or
ABCSparseSeries. This commit adds
ABCSparseDataFrame to that list -- so it
can detect that a DataFrame (which is an
array-like object) is sparse.

Added a test for this.

Author: sechilds

doc/source/whatsnew/v0.23.0.txt

@@ -987,6 +987,7 @@ Sparse
 - Bug in :class:`SparseDataFrame.to_csv` causing exception (:issue:`19384`)
 - Bug in :class:`SparseSeries.memory_usage` which caused segfault by accessing non sparse elements (:issue:`19368`)
 - Bug in constructing a ``SparseArray``: if ``data`` is a scalar and ``index`` is defined it will coerce to ``float64`` regardless of scalar's dtype. (:issue:`19163`)
+- Bug in :func:`is_sparse` which would report a ``SparseDataFrame`` as not sparse.
 
 Reshaping
 ^^^^^^^^^

pandas/core/dtypes/common.py

@@ -12,7 +12,8 @@
                      ExtensionDtype)
 from .generic import (ABCCategorical, ABCPeriodIndex,
                       ABCDatetimeIndex, ABCSeries,
-                      ABCSparseArray, ABCSparseSeries, ABCCategoricalIndex,
+                      ABCSparseArray, ABCSparseSeries, ABCSparseDataFrame,
+                      ABCCategoricalIndex,
                       ABCIndexClass, ABCDateOffset)
 from .inference import is_string_like, is_list_like
 from .inference import *  # noqa
@@ -149,7 +150,8 @@ def is_sparse(arr):
     False
     """
 
-    return isinstance(arr, (ABCSparseArray, ABCSparseSeries))
+    return isinstance(arr, (ABCSparseArray, ABCSparseSeries,
+                            ABCSparseDataFrame))
 
 
 def is_scipy_sparse(arr):

pandas/tests/dtypes/test_common.py

@@ -139,6 +139,7 @@ def test_is_object():
 def test_is_sparse(check_scipy):
     assert com.is_sparse(pd.SparseArray([1, 2, 3]))
     assert com.is_sparse(pd.SparseSeries([1, 2, 3]))
+    assert com.is_sparse(pd.SparseDataFrame([1, 2, 3]))
 
     assert not com.is_sparse(np.array([1, 2, 3]))