Question

我有一个2D文本文件：

[[1406], [1408], [1402], [1394, 102462], [1393], [20388], [20387, 20386], [1386], [1443, 1446, 766], [1432, 1438, 1430, 1416], [1442], [1434], [1430, 1416, 1417, 1419, 3446], [1429], [20011], [20015], [4435], [4441], [4443], [4444], [4448], [2433, 1413, 1418], [4450], [3444], [2478, 823, 3447], [3447], [2481, 1425, 942, 2476, 4449], [2482, 120, 3444], [13512], [3446], [13528]]

有什么方法可以将这个文件读入python吗？我尝试过：

from numpy import genfromtxt
con2 = genfromtxt('muti.txt', delimiter=',')
con2 = con2.astype(np.int64)

结果显示如下：

nan
nan
nan
nan
nan
nan
nan
1413.0
nan
nan
nan
nan
823.0
nan
nan
nan
1425.0
942.0
2476.0
nan
nan
120.0
nan
nan
nan
nan

数组中有很多nan。有人可以帮帮我吗？

Answer 1

我不知道numpy对此有什么功能，但由于您的文本文件恰好是有效的JSON，您可以将其加载为JSON，展平它，然后将结果转换为numpy数组。 / p>

>>> import json
>>> import numpy as np
>>> with open('muti.txt', 'r') as f: arr = json.load(f)
>>> np_arr = np.array([n for subarr in arr for n in subarr]).astype(np.int64)
>>> np_arr
array([  1406,   1408,   1402,   1394, 102462,   1393,  20388,  20387,
        20386,   1386,   1443,   1446,    766,   1432,   1438,   1430,
         1416,   1442,   1434,   1430,   1416,   1417,   1419,   3446,
         1429,  20011,  20015,   4435,   4441,   4443,   4444,   4448,
         2433,   1413,   1418,   4450,   3444,   2478,    823,   3447,
         3447,   2481,   1425,    942,   2476,   4449,   2482,    120,
         3444,  13512,   3446,  13528], dtype=int64)

Answer 2

如果这只是文本，那么只需替换/删除括号，拆分，转换为int。

txt = "[[1406], [1408], [1402], [1394, 102462], [1393], [20388], [20387, 20386], [1386], [1443, 1446, 766], [1432, 1438, 1430, 1416], [1442], [1434], [1430, 1416, 1417, 1419, 3446], [1429], [20011], [20015], [4435], [4441], [4443], [4444], [4448], [2433, 1413, 1418], [4450], [3444], [2478, 823, 3447], [3447], [2481, 1425, 942, 2476, 4449], [2482, 120, 3444], [13512], [3446], [13528]]"
data = [int(t) for t in txt.replace("]","").replace("[","").split(',')]

[1406,
1408,
1402,
1394,
...
13512,
3446,
13528]

python将2d读入1d数组

2 个答案: