Q&A
前提・実現したいこと
知人から頂いたコードで分からない事がありましたので、こちらにてお伺いさせて頂きます。
発生している問題・エラーメッセージ
python
1dists = [0.1, 35. , 54.4, ...] (1*n次元のndarrayです) 2idx = np.where(dists == np.min(dists))[0][0]
分からないこと
[0][0]というものが何を意味するのかが分かりません。
また，無くても正しくidxは出るのではないかと思っておりますので，何故使用する理由がよく分かりませんでした。
np.whereでネットで調べても出てこなかったのでこちらにて質問させていたきました。
よろしくお願いいたします。
補足情報（FW/ツールのバージョンなど）
python3
回答1件
0
こういう場合の調べ方です。
python
1>>> print(np.where.__doc__) 2 3 where(condition, [x, y]) 4 5 Return elements chosen from `x` or `y` depending on `condition`. 6 7 .. note:: 8 When only `condition` is provided, this function is a shorthand for 9 ``np.asarray(condition).nonzero()``. Using `nonzero` directly should be 10 preferred, as it behaves correctly for subclasses. The rest of this 11 documentation covers only the case where all three arguments are 12 provided. 13 14 Parameters 15 ---------- 16 condition : array_like, bool 17 Where True, yield `x`, otherwise yield `y`. 18 x, y : array_like 19 Values from which to choose. `x`, `y` and `condition` need to be 20 broadcastable to some shape. 21 22 Returns 23 ------- 24 out : ndarray 25 An array with elements from `x` where `condition` is True, and elements 26 from `y` elsewhere. 27 28 See Also 29 -------- 30 choose 31 nonzero : The function that is called when x and y are omitted 32 33 Notes 34 ----- 35 If all the arrays are 1-D, `where` is equivalent to:: 36 37 [xv if c else yv 38 for c, xv, yv in zip(condition, x, y)] 39 40 Examples 41 -------- 42 >>> a = np.arange(10) 43 >>> a 44 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 45 >>> np.where(a < 5, a, 10*a) 46 array([ 0, 1, 2, 3, 4, 50, 60, 70, 80, 90]) 47 48 This can be used on multidimensional arrays too: 49 50 >>> np.where([[True, False], [True, True]], 51 ... [[1, 2], [3, 4]], 52 ... [[9, 8], [7, 6]]) 53 array([[1, 8], 54 [3, 4]]) 55 56 The shapes of x, y, and the condition are broadcast together: 57 58 >>> x, y = np.ogrid[:3, :4] 59 >>> np.where(x < y, x, 10 + y) # both x and 10+y are broadcast 60 array([[10, 0, 0, 0], 61 [10, 11, 1, 1], 62 [10, 11, 12, 2]]) 63 64 >>> a = np.array([[0, 1, 2], 65 ... [0, 2, 4], 66 ... [0, 3, 6]]) 67 >>> np.where(a < 4, a, -1) # -1 is broadcast 68 array([[ 0, 1, 2], 69 [ 0, 2, -1], 70 [ 0, 3, -1]])
When only `condition` is provided, this function is a shorthand for ``np.asarray(condition).nonzero()``. Using `nonzero` directly should be preferred, as it behaves correctly for subclasses. The rest of this documentation covers only the case where all three arguments are provided.
なので、以下を実行します。
python
1>>> print(np.asarray(dists == np.min(dists)).nonzero.__doc__) 2a.nonzero() 3 4 Return the indices of the elements that are non-zero. 5 6 Refer to `numpy.nonzero` for full documentation. 7 8 See Also 9 -------- 10 numpy.nonzero : equivalent function
なので、以下を実行します。
python
1>>> print(np.nonzero.__doc__) 2 3 Return the indices of the elements that are non-zero. 4 5 Returns a tuple of arrays, one for each dimension of `a`, 6 containing the indices of the non-zero elements in that 7 dimension. The values in `a` are always tested and returned in 8 row-major, C-style order. 9 10 To group the indices by element, rather than dimension, use `argwhere`, 11 which returns a row for each non-zero element. 12 13 .. note:: 14 15 When called on a zero-d array or scalar, ``nonzero(a)`` is treated 16 as ``nonzero(atleast1d(a))``. 17 18 .. deprecated:: 1.17.0 19 20 Use `atleast1d` explicitly if this behavior is deliberate. 21 22 Parameters 23 ---------- 24 a : array_like 25 Input array. 26 27 Returns 28 ------- 29 tuple_of_arrays : tuple 30 Indices of elements that are non-zero. 31 32 See Also 33 -------- 34 flatnonzero : 35 Return indices that are non-zero in the flattened version of the input 36 array. 37 ndarray.nonzero : 38 Equivalent ndarray method. 39 count_nonzero : 40 Counts the number of non-zero elements in the input array. 41 42 Notes 43 ----- 44 While the nonzero values can be obtained with ``a[nonzero(a)]``, it is 45 recommended to use ``x[x.astype(bool)]`` or ``x[x != 0]`` instead, which 46 will correctly handle 0-d arrays. 47 48 Examples 49 -------- 50 >>> x = np.array([[3, 0, 0], [0, 4, 0], [5, 6, 0]]) 51 >>> x 52 array([[3, 0, 0], 53 [0, 4, 0], 54 [5, 6, 0]]) 55 >>> np.nonzero(x) 56 (array([0, 1, 2, 2]), array([0, 1, 0, 1])) 57 58 >>> x[np.nonzero(x)] 59 array([3, 4, 5, 6]) 60 >>> np.transpose(np.nonzero(x)) 61 array([[0, 0], 62 [1, 1], 63 [2, 0], 64 [2, 1]]) 65 66 A common use for ``nonzero`` is to find the indices of an array, where 67 a condition is True. Given an array `a`, the condition `a` > 3 is a 68 boolean array and since False is interpreted as 0, np.nonzero(a > 3) 69 yields the indices of the `a` where the condition is true. 70 71 >>> a = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]) 72 >>> a > 3 73 array([[False, False, False], 74 [ True, True, True], 75 [ True, True, True]]) 76 >>> np.nonzero(a > 3) 77 (array([1, 1, 1, 2, 2, 2]), array([0, 1, 2, 0, 1, 2])) 78 79 Using this result to index `a` is equivalent to using the mask directly: 80 81 >>> a[np.nonzero(a > 3)] 82 array([4, 5, 6, 7, 8, 9]) 83 >>> a[a > 3] # prefer this spelling 84 array([4, 5, 6, 7, 8, 9]) 85 86 ``nonzero`` can also be called as a method of the array. 87 88 >>> (a > 3).nonzero() 89 (array([1, 1, 1, 2, 2, 2]), array([0, 1, 2, 0, 1, 2]))
で、
Returns a tuple of arrays, one for each dimension of `a`, containing the indices of the non-zero elements in that dimension. The values in `a` are always tested and returned in row-major, C-style order.
なので、[0][0]が必要です。
投稿2021/06/25 15:33
総合スコア24670
あなたの回答
tips
プレビュー
バッドをするには、ログインかつ
こちらの条件を満たす必要があります。
2021/06/26 03:04
2021/06/28 02:57 編集