Return (maximum - minimum) along the given dimension (i.e. peak-to-peak value). .. warning:: `ptp` preserves the data type of the array. This means the return value for an input of signed integers with n bits (e.g. `np.int8`, `np.int16`,
(self, axis=None, out=None, fill_value=None, keepdims=False)
| 5964 | return out |
| 5965 | |
| 5966 | def ptp(self, axis=None, out=None, fill_value=None, keepdims=False): |
| 5967 | """ |
| 5968 | Return (maximum - minimum) along the given dimension |
| 5969 | (i.e. peak-to-peak value). |
| 5970 | |
| 5971 | .. warning:: |
| 5972 | `ptp` preserves the data type of the array. This means the |
| 5973 | return value for an input of signed integers with n bits |
| 5974 | (e.g. `np.int8`, `np.int16`, etc) is also a signed integer |
| 5975 | with n bits. In that case, peak-to-peak values greater than |
| 5976 | ``2**(n-1)-1`` will be returned as negative values. An example |
| 5977 | with a work-around is shown below. |
| 5978 | |
| 5979 | Parameters |
| 5980 | ---------- |
| 5981 | axis : {None, int}, optional |
| 5982 | Axis along which to find the peaks. If None (default) the |
| 5983 | flattened array is used. |
| 5984 | out : {None, array_like}, optional |
| 5985 | Alternative output array in which to place the result. It must |
| 5986 | have the same shape and buffer length as the expected output |
| 5987 | but the type will be cast if necessary. |
| 5988 | fill_value : scalar or None, optional |
| 5989 | Value used to fill in the masked values. |
| 5990 | keepdims : bool, optional |
| 5991 | If this is set to True, the axes which are reduced are left |
| 5992 | in the result as dimensions with size one. With this option, |
| 5993 | the result will broadcast correctly against the array. |
| 5994 | |
| 5995 | Returns |
| 5996 | ------- |
| 5997 | ptp : ndarray. |
| 5998 | A new array holding the result, unless ``out`` was |
| 5999 | specified, in which case a reference to ``out`` is returned. |
| 6000 | |
| 6001 | Examples |
| 6002 | -------- |
| 6003 | >>> x = np.ma.MaskedArray([[4, 9, 2, 10], |
| 6004 | ... [6, 9, 7, 12]]) |
| 6005 | |
| 6006 | >>> x.ptp(axis=1) |
| 6007 | masked_array(data=[8, 6], |
| 6008 | mask=False, |
| 6009 | fill_value=999999) |
| 6010 | |
| 6011 | >>> x.ptp(axis=0) |
| 6012 | masked_array(data=[2, 0, 5, 2], |
| 6013 | mask=False, |
| 6014 | fill_value=999999) |
| 6015 | |
| 6016 | >>> x.ptp() |
| 6017 | 10 |
| 6018 | |
| 6019 | This example shows that a negative value can be returned when |
| 6020 | the input is an array of signed integers. |
| 6021 | |
| 6022 | >>> y = np.ma.MaskedArray([[1, 127], |
| 6023 | ... [0, 127], |