Question

您好我正在尝试将4x4矩阵从字典加载到np数组中，

u'{"column0":{"x":1,"y":0.0,"z":0.0,"w":-1.0},
   "column1":{"x":0.0,"y":1,"z":0.0,"w":-1.0},
   "column2":{"x":1,"y":1,"z":1.0,"w":-1.0},
   "column3":{"x":-1.0,"y":-1.0,"z":-1.0,"w":-1.0}
  }'

我想将其加载为

array([[ 1,0., 0.,-1 ],
       [ 0.,1,0.,-1 ],
       [ 1,1,1,-1 ],
       [ -1,-1,-1,-1 ]])

目前我正在遍历每个元素，我确信有更好的pythonic方式直接使用pandas或numpy。有什么建议吗？

Answer 1

您可以通过numpy或pandas执行此操作。这两种方法都需要ast.literal_eval。

由于字典是无序的，因此每种方法都需要注意确保按需要对列进行排序。

<强>设置

import numpy as np, pandas as pd
from ast import literal_eval

mystr = """{"column0":{"x":1,"y":0.0,"z":0.0,"w":-1.0},
   "column1":{"x":0.0,"y":1,"z":0.0,"w":-1.0},
   "column2":{"x":1,"y":1,"z":1.0,"w":-1.0},
   "column3":{"x":-1.0,"y":-1.0,"z":-1.0,"w":-1.0}
  }"""

d = literal_eval(mystr)

选项1：numpy

res_np = np.array([[v[j] for j in ['x', 'y', 'z', 'w']] for k, v in d.items()])

# array([[ 1.,  0.,  0., -1.],
#        [ 0.,  1.,  0., -1.],
#        [ 1.,  1.,  1., -1.],
#        [-1., -1., -1., -1.]])

选项2：pandas

res_df = pd.DataFrame(d).T.iloc[:, [1, 2, 3, 0]].values

# array([[ 1.,  0.,  0., -1.],
#        [ 0.,  1.,  0., -1.],
#        [ 1.,  1.,  1., -1.],
#        [-1., -1., -1., -1.]])

Answer 2

您可以使用ast.literal_eval评估字典字符串，将其放入pandas转置数据框（x，y，z，{{1按照您的列顺序排列），然后使用w：

as_matrix

如何在python

2 个答案: