Sponsored Link

(70) numpy.reshapeは浅いコピー

(70) numpy.reshapeは浅いコピー

4行3列の乱数配列aを作成する。

>>> import numpy as np
>>> a = np.random.rand(4,3)
>>> a
array([[ 0.15940847,  0.56564422,  0.82868468],
       [ 0.55344639,  0.55527501,  0.75242168],
       [ 0.43876864,  0.50025427,  0.07168502],
       [ 0.15186   ,  0.00686738,  0.98197383]])

配列aを numpy.reshape で 6行2列に変形する。

>>> b = np.reshape(a, (6,2))
>>> b
array([[ 0.15940847,  0.56564422],
       [ 0.82868468,  0.55344639],
       [ 0.55527501,  0.75242168],
       [ 0.43876864,  0.50025427],
       [ 0.07168502,  0.15186   ],
       [ 0.00686738,  0.98197383]])

配列bの要素の値を書き換える。

>>> b[3,1] = 0
>>> b
array([[ 0.15940847,  0.56564422],
       [ 0.82868468,  0.55344639],
       [ 0.55527501,  0.75242168],
       [ 0.43876864,  0.        ],
       [ 0.07168502,  0.15186   ],
       [ 0.00686738,  0.98197383]])

すると、配列aの要素も書き換わっている!

>>> a
array([[ 0.15940847,  0.56564422,  0.82868468],
       [ 0.55344639,  0.55527501,  0.75242168],
       [ 0.43876864,  0.        ,  0.07168502],
       [ 0.15186   ,  0.00686738,  0.98197383]])

numpy.reshape したものを別のオブジェクトに代入したい場合、numpy.copy を使用する。

>>> a = np.random.rand(4,3)
>>> a
array([[ 0.59587499,  0.33831817,  0.75324048],
       [ 0.67519994,  0.57802629,  0.48760045],
       [ 0.27385989,  0.02654679,  0.46217751],
       [ 0.20590104,  0.31475441,  0.31250744]])
>>> b = a.reshape(6,2).copy()
>>> b
array([[ 0.59587499,  0.33831817],
       [ 0.75324048,  0.67519994],
       [ 0.57802629,  0.48760045],
       [ 0.27385989,  0.02654679],
       [ 0.46217751,  0.20590104],
       [ 0.31475441,  0.31250744]])
>>> b[3,0] = 0
>>> b
array([[ 0.59587499,  0.33831817],
       [ 0.75324048,  0.67519994],
       [ 0.57802629,  0.48760045],
       [ 0.        ,  0.02654679],
       [ 0.46217751,  0.20590104],
       [ 0.31475441,  0.31250744]])
>>> a
array([[ 0.59587499,  0.33831817,  0.75324048],
       [ 0.67519994,  0.57802629,  0.48760045],
       [ 0.27385989,  0.02654679,  0.46217751],
       [ 0.20590104,  0.31475441,  0.31250744]])

こう書いても同じ。

>>> a = np.random.rand(4,3)
>>> a
array([[ 0.59738729,  0.8289699 ,  0.22344194],
       [ 0.86271615,  0.76692546,  0.1131666 ],
       [ 0.61305503,  0.28302821,  0.24619501],
       [ 0.10828006,  0.38060943,  0.46060143]])
>>> b = np.copy(a.reshape(6,2))
>>> b
array([[ 0.59738729,  0.8289699 ],
       [ 0.22344194,  0.86271615],
       [ 0.76692546,  0.1131666 ],
       [ 0.61305503,  0.28302821],
       [ 0.24619501,  0.10828006],
       [ 0.38060943,  0.46060143]])
>>> b[3,0] = 0
>>> b
array([[ 0.59738729,  0.8289699 ],
       [ 0.22344194,  0.86271615],
       [ 0.76692546,  0.1131666 ],
       [ 0.        ,  0.28302821],
       [ 0.24619501,  0.10828006],
       [ 0.38060943,  0.46060143]])
>>> a
array([[ 0.59738729,  0.8289699 ,  0.22344194],
       [ 0.86271615,  0.76692546,  0.1131666 ],
       [ 0.61305503,  0.28302821,  0.24619501],
       [ 0.10828006,  0.38060943,  0.46060143]])
Sponsored Link

コメントを残す

メールアドレスが公開されることはありません。 * が付いている欄は必須項目です

*