如何创建保留列的布尔掩码？

Question

例如：说我有

x = np.array[['positive','a'],['negative','b'],['positive','c'], ['negative','d']]

我想要一个值为

的新数组

v = [['positive','a'], ['positive','c']]

这是怎么做到的？谢谢你的时间！

Answer 1

更正语法：

In [359]: x = np.array([['positive','a'],['negative','b'],['positive','c'], ['ne
     ...: gative','d']])
In [360]: x
Out[360]: 
array([['positive', 'a'],
       ['negative', 'b'],
       ['positive', 'c'],
       ['negative', 'd']], dtype='<U8')

只有 select 个带有索引的“行”：

In [361]: x[[0,2],:]
Out[361]: 
array([['positive', 'a'],
       ['positive', 'c']], dtype='<U8')

等效的布尔掩码：

In [362]: mask=np.zeros(4,bool)
In [363]: mask[[0,2]]=True
In [364]: mask
Out[364]: array([ True, False,  True, False])
In [365]: x[mask,:]
Out[365]: 
array([['positive', 'a'],
       ['positive', 'c']], dtype='<U8')

如何创建保留列的布尔掩码？

How to create a Boolean Mask which preserves a column?

python

arrays

boolean

numpy

mask