Coverage for qdscreen/tests/compat.py: 55%

1import numpy as np 

2import pandas as pd 

3 

4 

5def patch_pandas_if_needed(): 

6 try: 

7 pd.DataFrame.to_numpy 

8 except AttributeError: 

9 def to_numpy(self, dtype=None, copy=False): 

10 """ 

11 Convert the DataFrame to a NumPy array. 

12 

13 .. versionadded:: 0.24.0 

14 

15 By default, the dtype of the returned array will be the common NumPy 

16 dtype of all types in the DataFrame. For example, if the dtypes are 

17 ``float16`` and ``float32``, the results dtype will be ``float32``. 

18 This may require copying data and coercing values, which may be 

19 expensive. 

20 

21 Parameters 

22 ---------- 

23 dtype : str or numpy.dtype, optional 

24 The dtype to pass to :meth:`numpy.asarray` 

25 copy : bool, default False 

26 Whether to ensure that the returned value is a not a view on 

27 another array. Note that ``copy=False`` does not *ensure* that 

28 ``to_numpy()`` is no-copy. Rather, ``copy=True`` ensure that 

29 a copy is made, even if not strictly necessary. 

30 

31 Returns 

32 ------- 

33 numpy.ndarray 

34 

35 See Also 

36 -------- 

37 Series.to_numpy : Similar method for Series. 

38 

39 Examples 

40 -------- 

41 >>> pd.DataFrame({"A": [1, 2], "B": [3, 4]}).to_numpy() 

42 array([[1, 3], 

43 [2, 4]]) 

44 

45 With heterogenous data, the lowest common type will have to 

46 be used. 

47 

48 >>> df = pd.DataFrame({"A": [1, 2], "B": [3.0, 4.5]}) 

49 >>> df.to_numpy() 

50 array([[1. , 3. ], 

51 [2. , 4.5]]) 

52 

53 For a mix of numeric and non-numeric types, the output array will 

54 have object dtype. 

55 

56 >>> df['C'] = pd.date_range('2000', periods=2) 

57 >>> df.to_numpy() 

58 array([[1, 3.0, Timestamp('2000-01-01 00:00:00')], 

59 [2, 4.5, Timestamp('2000-01-02 00:00:00')]], dtype=object) 

60 """ 

61 result = np.array(self.values, dtype=dtype, copy=copy) 

62 return result 

63 

64 pd.DataFrame.to_numpy = to_numpy 

65 

66 else: 

67 pass