numpy

Tweet

特定の列でソートする

argsortを使うのがポイント。

# coding:UTF-8
import numpy as np

if __name__ == '__main__':
    data = np.random.randn(5, 6)
    print(data)
    col_num = 4

    print(data[:, col_num])

    col = np.argsort(data[:, col_num])
    print(col)

    sorted_col = data[np.argsort(data[:, col_num])]
    print(sorted_col)

実行結果

[[-0.14016536  0.21551254  0.24784852  0.59144645  0.32235863  0.67189879]
 [-0.1197581   0.34133468  0.89260091 -1.09935213  1.52181822  0.01690957]
 [-0.88335896 -0.72368754 -0.87826505  2.48254012 -0.62577154  1.82120588]
 [ 0.46286874  1.74054554  1.21514558  0.520769    1.01164945 -0.57376935]
 [-2.14264006 -0.36363893  1.18613253  0.04081303 -0.22348603 -0.26786658]]
[ 0.32235863  1.52181822 -0.62577154  1.01164945 -0.22348603]
[2 4 0 3 1]
[[-0.88335896 -0.72368754 -0.87826505  2.48254012 -0.62577154  1.82120588]
 [-2.14264006 -0.36363893  1.18613253  0.04081303 -0.22348603 -0.26786658]
 [-0.14016536  0.21551254  0.24784852  0.59144645  0.32235863  0.67189879]
 [ 0.46286874  1.74054554  1.21514558  0.520769    1.01164945 -0.57376935]
 [-0.1197581   0.34133468  0.89260091 -1.09935213  1.52181822  0.01690957]]

2次元配列の行をランダムに入れ替える

https://hahaeatora.hateblo.jp/entry/2019/01/29/200000を参考に。