将numpy数组转换为json答案

【问题标题】：convert numpy array to json将numpy数组转换为json
【发布时间】：2020-10-08 07:02:03
【问题描述】：

这里是输入类型，一个NumPy数组，比如

[['id' '796132' '1148512' '87741691' '87910849' '88188296']
 ['diagnosis' '1' '0' '1' '0' '0']
 ['fractal_dimension_worst' '0.08701' '0.08006' '0.08255' '0.06484' '0.1118']]

我想把它转换成一个 JSON 文件，比如

{
   "id": ['796132', '1148512', '87741691', '87910849', '88188296'],
   "diagnosis": ['1', '0', '1', '0', '0'],
   "fractal_dimension_worst": ['0.08701', '0.08006', '0.08255', '0.06484', '0.1118']
}

怎么做？

【问题讨论】：

只要把它当作一个列表的列表，即使是arr.tolist()。然后从中创建字典。应该是一个直接的迭代。

标签： python arrays json numpy

【解决方案1】：

您可以像这样使用字典推导：

import numpy as np
arr = np.array([
    ['id', '796132', '1148512', '87741691', '87910849', '88188296'],
    ['diagnosis', '1', '0', '1', '0', '0'],
    ['fractal_dimension_worst', '0.08701', '0.08006', '0.08255', '0.06484', '0.1118']
])

json_dict = {l[0]: list(l[1:]) for l in arr}

>> {'id': ['796132', '1148512', '87741691', '87910849', '88188296'],
 'diagnosis': ['1', '0', '1', '0', '0'],
 'fractal_dimension_worst': ['0.08701',
  '0.08006',
  '0.08255',
  '0.06484',
  '0.1118']}

【讨论】：

【解决方案2】：

在受到@hpaulj 的启发后，我就是这样做的。

import csv
import numpy as np
import pandas as pd

# df = pd.read_csv('inf_small.csv')

with open('..\inf.csv', newline='') as csvfile:
    data = list(csv.reader(csvfile, delimiter=','))

npMatrix = np.array(data)

# calculate row and column numbers
row_count, column_count = npMatrix.shape
# neglect first row and get new row numbers
row_count = row_count - 1
npMatrix = npMatrix.transpose()

# transfer numpy array to list
matrix = npMatrix.tolist()

# transfer list to that JSON file
result = {}

for index, item in enumerate(matrix):
    temp = item[0]
    del item[0]
    result[temp] = item

【讨论】：