How to modify values of a column of a 2D tensor based on condition - Tensorflow?

Question

I have a 2D tensor and want values of its last column to be 0 if values > 0 and 1 otherwise. It should behave somewhat similar to the following block of numpy code:

x = np.random.rand(8, 4)
x[:, -1] = np.where(x[:, -1] > 0, 0, 1)

Is there a way to achieve the same behavior for a 2D tensor in Tensorflow?

Answer 1

This might not be the most elegant solution, but it works:

x=tf.ones((5,10))

rows=tf.stack(tf.range(tf.shape(x)[0]))
column=tf.ones_like(rows)*tf.shape(x)[1]-1
idx=tf.stack((rows,column),axis=1)
x_new=tf.tensor_scatter_nd_update(x, idx, tf.where(x[:, -1] > 0, 0., 1.))

print(x_new)

And the result looks like this (the original x is a tf.ones):

[[1. 1. 1. 1. 1. 1. 1. 1. 1. 0.]
 [1. 1. 1. 1. 1. 1. 1. 1. 1. 0.]
 [1. 1. 1. 1. 1. 1. 1. 1. 1. 0.]
 [1. 1. 1. 1. 1. 1. 1. 1. 1. 0.]
 [1. 1. 1. 1. 1. 1. 1. 1. 1. 0.]]