add sort_labels performance tip in docstrings

pandas/core/arrays/sparse/accessor.py

@@ -113,6 +113,8 @@ def to_coo(self, row_levels=(0,), column_levels=(1,), sort_labels=False):
  column_levels : tuple/list
  sort_labels : bool, default False
  Sort the row and column labels before forming the sparse matrix.
+ When `row_levels` and/or `column_levels` refer to a single level,
+ set to `True` for a faster execution.
 
  Returns
  -------

pandas/core/arrays/sparse/scipy_sparse.py

@@ -54,7 +54,8 @@ def _levels_to_axis(
  valid_ilocs : numpy.ndarray
  Array of integer positions of valid values for the sparse matrix in ss.
  sort_labels : bool, default False
- Sort the axis labels before forming the sparse matrix.
+ Sort the axis labels before forming the sparse matrix. When `levels`
+ refers to a single level, set to True for a faster execution.
 
  Returns
  -------
@@ -104,6 +105,8 @@ def _to_ijv(
  column_levels : tuple/list
  sort_labels : bool, default False
  Sort the row and column labels before forming the sparse matrix.
+ When `row_levels` and/or `column_levels` refer to a single level,
+ set to `True` for a faster execution.
 
  Returns
  -------