将多维数组字符串转换为python中的浮点数答案

【问题标题】：Convert Multidimensional array string to float in python将多维数组字符串转换为python中的浮点数
【发布时间】：2015-10-20 03:12:24
【问题描述】：

我有一个这样的多维数组：

array([['120', '29.9475077984', '1'],
       ['125', '31.3887667742', '1'],
       ['125', '32.3881706091', '1'],
       ['125', '34.4894481007', '1'],
       ['126', '36.1494551046', '1'],
       ['127', '39.3121447948', '1'],
       ['128', '43.8203811171', '1'],
       ['128', '49.3179066095', '1'],
       ['128', '53.4929489926', '1'],
       ['128', '55.1837748899', '1'],
       ['130', '55.9167038553', '1'],
       ['130', '56.2727376481', '1'],
       ['130', '57.480058071', '1'],
       ['130', '60.3922465138', '1'],
       ['130', '61.2506277637', '1'],
       ['130', '60.5279054759', '1'],
       ['143', '62.139526711', '1'],
       ['143', '65.4147315349', '1'],
       ['143', '72.3278873965', '1'],

如您所见，所有值都是字符串。我需要将它们转换为浮点值，有办法吗？我找到了转换单个字符串的解决方案，但它不适用于数组。

【问题讨论】：

看起来像ndarray，调用astype(float)就可以了。
我添加了numpy标签。

标签： python string numpy multidimensional-array integer

【解决方案1】：

您可以在列表理解中使用map：

>>> li=[['120', '29.9475077984', '1'],
...        ['125', '31.3887667742', '1'],
...        ['125', '32.3881706091', '1'],
...        ['125', '34.4894481007', '1'],
...        ['126', '36.1494551046', '1'],
...        ['127', '39.3121447948', '1'],
...        ['128', '43.8203811171', '1'],
...        ['128', '49.3179066095', '1'],
...        ['128', '53.4929489926', '1'],
...        ['128', '55.1837748899', '1'],
...        ['130', '55.9167038553', '1'],
...        ['130', '56.2727376481', '1'],
...        ['130', '57.480058071', '1'],
...        ['130', '60.3922465138', '1'],
...        ['130', '61.2506277637', '1'],
...        ['130', '60.5279054759', '1'],
...        ['143', '62.139526711', '1'],
...        ['143', '65.4147315349', '1'],
...        ['143', '72.3278873965', '1']]
>>> 
>>> 
>>> [map(float,i) for i in li]
[[120.0, 29.9475077984, 1.0], [125.0, 31.3887667742, 1.0], [125.0, 32.3881706091, 1.0], [125.0, 34.4894481007, 1.0], [126.0, 36.1494551046, 1.0], [127.0, 39.3121447948, 1.0], [128.0, 43.8203811171, 1.0], [128.0, 49.3179066095, 1.0], [128.0, 53.4929489926, 1.0], [128.0, 55.1837748899, 1.0], [130.0, 55.9167038553, 1.0], [130.0, 56.2727376481, 1.0], [130.0, 57.480058071, 1.0], [130.0, 60.3922465138, 1.0], [130.0, 61.2506277637, 1.0], [130.0, 60.5279054759, 1.0], [143.0, 62.139526711, 1.0], [143.0, 65.4147315349, 1.0], [143.0, 72.3278873965, 1.0]]

如果你有一个 numpy 数组，你可以使用 np.astype ：

>>> li.astype(float)
array([[ 120.        ,   29.9475078 ,    1.        ],
       [ 125.        ,   31.38876677,    1.        ],
       [ 125.        ,   32.38817061,    1.        ],
       [ 125.        ,   34.4894481 ,    1.        ],
       [ 126.        ,   36.1494551 ,    1.        ],
       [ 127.        ,   39.31214479,    1.        ],
       [ 128.        ,   43.82038112,    1.        ],
       [ 128.        ,   49.31790661,    1.        ],
       [ 128.        ,   53.49294899,    1.        ],
       [ 128.        ,   55.18377489,    1.        ],
       [ 130.        ,   55.91670386,    1.        ],
       [ 130.        ,   56.27273765,    1.        ],
       [ 130.        ,   57.48005807,    1.        ],
       [ 130.        ,   60.39224651,    1.        ],
       [ 130.        ,   61.25062776,    1.        ],
       [ 130.        ,   60.52790548,    1.        ],
       [ 143.        ,   62.13952671,    1.        ],
       [ 143.        ,   65.41473153,    1.        ],
       [ 143.        ,   72.3278874 ,    1.        ]])

【讨论】：

【解决方案2】：

你可以使用np.ndarray.astype()函数-

例子-

In [5]: n= np.array([['120', '29.9475077984', '1'],
   ...:        ['125', '31.3887667742', '1'],
   ...:        ['125', '32.3881706091', '1'],
   ...:        ['125', '34.4894481007', '1'],
   ...:        ['126', '36.1494551046', '1'],
   ...:        ['127', '39.3121447948', '1'],
   ...:        ['128', '43.8203811171', '1'],
   ...:        ['128', '49.3179066095', '1'],
   ...:        ['128', '53.4929489926', '1'],
   ...:        ['128', '55.1837748899', '1'],
   ...:        ['130', '55.9167038553', '1'],
   ...:        ['130', '56.2727376481', '1'],
   ...:        ['130', '57.480058071', '1'],
   ...:        ['130', '60.3922465138', '1'],
   ...:        ['130', '61.2506277637', '1'],
   ...:        ['130', '60.5279054759', '1'],
   ...:        ['143', '62.139526711', '1'],
   ...:        ['143', '65.4147315349', '1'],
   ...:        ['143', '72.3278873965', '1']])

In [8]: n = n.astype(np.float)

In [9]: n
Out[9]:
array([[ 120.        ,   29.9475078 ,    1.        ],
       [ 125.        ,   31.38876677,    1.        ],
       [ 125.        ,   32.38817061,    1.        ],
       [ 125.        ,   34.4894481 ,    1.        ],
       [ 126.        ,   36.1494551 ,    1.        ],
       [ 127.        ,   39.31214479,    1.        ],
       [ 128.        ,   43.82038112,    1.        ],
       [ 128.        ,   49.31790661,    1.        ],
       [ 128.        ,   53.49294899,    1.        ],
       [ 128.        ,   55.18377489,    1.        ],
       [ 130.        ,   55.91670386,    1.        ],
       [ 130.        ,   56.27273765,    1.        ],
       [ 130.        ,   57.48005807,    1.        ],
       [ 130.        ,   60.39224651,    1.        ],
       [ 130.        ,   61.25062776,    1.        ],
       [ 130.        ,   60.52790548,    1.        ],
       [ 143.        ,   62.13952671,    1.        ],
       [ 143.        ,   65.41473153,    1.        ],
       [ 143.        ,   72.3278874 ,    1.        ]])

【讨论】：

感谢您的回答。我刚刚尝试过你的方法，但它返回： ValueError Traceback (most recent call last) in () ----> 1 n=n.astype(np.float ) ValueError: 无法将字符串转换为浮点数：
您有一些不能转换为浮点数的字符串？检查您的数组，看看是否有任何字符串无法转换为字符串。