我有一个pandas
系列对象,如下所示:
s1 = pd.Series([0,1,2,3,4,5,6,7,8], index=['AB', 'AC','AD', 'BA','BB','BC','CA','CB','CC'])
我要将该系列转换为numpy
数组,如下所示:
series_size = s1.size
dimension_len = np.sqrt(series_size)
**Note: series_size will always have an integer sqrt
dimension_len将确定所需的二维数组中每个维的大小。
在上述系列对象中,Dimension_len = 3,因此所需的numpy
数组将是3 x 3的数组,如下所示:
np.array([[0, 1, 2],
[3, 4, 5],
[6,7, 8]])
我有一个pandas
数据框对象,如下所示:
s1 = pd.Series([0,1,2,3,4,5,6,7,8], index=['AA', 'AB','AC', 'BA','BB','BC','CA','CB','CC'])
s2 = pd.Series([-2,2], index=['AB','BA'])
s3 = pd.Series([4,3,-3,-4], index=['AC','BC', 'CB','CA'])
df = pd.concat([s1, s2, s3], axis=1)
max_size = max(s1.size, s2.size, s3.size)
dimension_len = np.sqrt(max_size)
num_columns = len(df.columns)
**Note: max_size will always have an integer sqrt
结果numpy
数组将由以下信息确定:
num_columns =确定数组的维数 Dimensions_len =确定每个尺寸的大小
在上面的示例中,所需的numpy
数组将为3 x 3 x 3(num_columns = 3并且Dimension_len = 3)
同样,df的第一列将变为DESIRED_ARRAY [0],df的第二列将变为DESIRED_ARRAY [1],df的第三列将变为DESIRED_ARRAY [2],依此类推...
我想要的所需数组如下:
np.array([[[0, 1, 2],
[3, 4, 5],
[6, 7, 8]],
[[np.nan,-2, np.nan],
[2, np.nan, np.nan],
[np.nan, np.nan, np.nan]],
[[np.nan,np.nan, 4],
[np.nan, np.nan, 3],
[-4, -3, np.nan]],
])
答案 0 :(得分:1)
IIUC,您可以尝试numpy转置和reshape
df.values.T.reshape(-1, int(dimension_len), int(dimension_len))
Out[30]:
array([[[ 0., 1., 2.],
[ 3., 4., 5.],
[ 6., 7., 8.]],
[[nan, -2., nan],
[ 2., nan, nan],
[nan, nan, nan]],
[[nan, nan, 4.],
[nan, nan, 3.],
[-4., -3., nan]]])