根据每个数组的特定索引中的值，在 2D numpy 数组中按索引查找和切片？

Question

I have many 2D numpy arrays that look like this:

arr = np.array([[2,2],
                [2,3],
                [3,4],
                [3,5],
                [3,6],
                [4,7]))

How can I query the 2D array and retrieve all the arrays with a value of, for example, 3 in the 0 index? So I would want:

[[3,4],[3,5],[3,6]]

I considered turning it into a list of lists, but that seems inefficient as I have a lot of queries to make. Using np.argwhere or np.where doesn't seem to isolate by index value, either.

Thank you.

Answer 1

For advanced indexing, at first we must find elements which have the specified number as its first argument by arr[:, 0] == 3 , so:

arr[arr[:, 0] == 3]