将浮点字符串转换为不同行的numpy数组

Question

我有一个字符串 final_line，内容是：

0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0130631958731,0.0,0.0,0.0,0.0,0.0,0.0,0.0130631958731,0.0,0.0,0.0,0.0,0.0,0.0,,
0.00507937313707,0.0,0.0,0.0201058520009,0.0,0.0,0.0459562331449,0.0268078026679,0.0,0.0103772139359,0.0,0.0,0.0438673134565,0.0,0.0268078026679,0.0,0.0,0.0,0.0,0.0,0.0224437417685,,
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.070802847389,0.0,0.0,,
0.0,0.0,0.169140135429,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0261263917462,0.0,0.0,0.0,0.0,0.0,0.0,0.0261263917462,0.0,0.0,0.0,0.0,0.0,0.0,,
0.0961428138228,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,

我想把它转换成 numpy 数组，而不需要先转成 CSV 文件。在这个字符串中，每一行都应该变成数组中的一行，像这样：

[['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0.08364751'], ['0.017944717', '0', '0', '0', '0.009470823', '0', '0', '0', '0', '0', '0'], ['0.012620501', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0'], ['0.012620501', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0'], ['0', '0', '0', '0', '0.01332164', '0', '0', '0', '0', '0', '0'], ['0.012620501', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0'], ['0', '0', '0', '0', '0', '0.097414177', '0.042092545', '0', '0', '0', '0'], ['0', '0', '0', '0', '0.01332164', '0', '0', '0', '0', '0', '0'], ['0.005324215', '0', '0', '0.04598186', '0.028100025', '0', '0', '0', '0', '0', '0.023525603'], ['0', '0', '0', '0', '0', '0', '0', '0', '0.055765006', '0', '0'], ['0', '0', '0.133216404', '0', '0', '0', '0', '0', '0', '0', '0']]

我知道输出的值可能不同，但我只是为了展示格式。

举个例子：

我有这样的字符串：

a1,b1,c1,d1,e1,,
a2,b2,c2,d2,e2,,
a3,b3,c3,d3,e3,,

那么输出应该是：

[['a1','b1','c1','d1','e1'],
 ['a2','b2','c2','d2','e2'],
 ['a3','b3','c3','d3','e3']]

我看到过其他问题，但它们只把字符串转换成了数组的一行。我想要的是每一行都变成数组中的新一行，正如例子所示。

Answer 1

你的数据是以列表的形式存在的，每一行之间用一个额外的逗号分隔。

d_in = "1,2,3,,4,5,6,,7,8,9"
first = d_in.split(",,")
final = [line.split(",") for line in first]

first 用来存储一组字符串（也就是你的每一列）。final 则用来把你的数据整理成正确的格式（在我的例子中，格式是 [[1,2,3],[4,5,6],[7,8,9]]）

Answer 2

map 在这种情况下可以很有用。另外，你的 final list 最后的那个 ,, 是多余的，所以我就忽略掉了（在进行 split 操作之前，看看 :-2）。

import numpy as np

A = final_line[:-2].split(',,')
B = np.array([map(float,a.split(',')) for a in A])