使用numpy获取切片后剩余的内容

Question

I have a numpy matrix where one row for example looks like the following:

|0 1 2 3 4 5 6 7 8|

I can get a certain piece of the array eg. [3,4,5] which I need for one purpose using numpy slicing a[0,3:6] .

Is there anything builtin that will allow me to also get everything not in that range with it? Like [0,1,2,6,7,8]

Answer 1

One approach with boolean indexing -

a[~np.in1d(np.arange(a.size),r)]

Sample run -

In [174]: a
Out[174]: array([10, 11, 12, 13, 14, 15, 16, 17, 18])

In [175]: r
Out[175]: [3, 4, 5]

In [176]: a[~np.in1d(np.arange(a.size),r)]  # Without r
Out[176]: array([10, 11, 12, 16, 17, 18])

In [177]: a[r]  # With r
Out[177]: array([13, 14, 15])

Another with integer array indexing -

a[np.setdiff1d(np.arange(a.size),r)]

Another way would be concatenating slices on either sides of the original slice -

np.concatenate((a[:r[0]], a[r[-1]+1:]))

Answer 2

There's some ambiguity in your question and example. Are you selecting elements by value or index? And should we take slice literally?

Taking slice literally:

In [10]: x=np.arange(10)   # stick with the ambiguous input for now
In [11]: x[3:6]
Out[11]: array([3, 4, 5])

np.delete is a handy tool if selecting elements by position. It's general purpose, and can use slice as well as list inputs:

In [13]: np.delete(x,slice(3,6))
Out[13]: array([0, 1, 2, 6, 7, 8, 9])
In [14]: np.delete(x,[3,4,5])
Out[14]: array([0, 1, 2, 6, 7, 8, 9])

np.in1d is useful if you want to select elements by value.

Boolean masking is also a good tool to know and use.

---

delete uses different methods depending on the inputs. For a simple slice I believe it uses the equivalent of:

In [15]: np.concatenate((x[:3],x[6:]))
Out[15]: array([0, 1, 2, 6, 7, 8, 9])