ML Community Day is November 9! Join us for updates from TensorFlow, JAX, and more Learn more


Return the elements where condition is True (multiplexing x and y).

Used in the notebooks

Used in the guide Used in the tutorials

This operator has two modes: in one mode both x and y are provided, in another mode neither are provided. condition is always expected to be a tf.Tensor of type bool.

Retrieving indices of True elements

If x and y are not provided (both are None):

tf.where will return the indices of condition that are True, in the form of a 2-D tensor with shape (n, d). (Where n is the number of matching indices in condition, and d is the number of dimensions in condition).

Indices are output in row-major order.

tf.where([True, False, False, True])
<tf.Tensor: shape=(2, 1), dtype=int64, numpy=
tf.where([[True, False], [False, True]])
<tf.Tensor: shape=(2, 2), dtype=int64, numpy=
array([[0, 0],
       [1, 1]])>
tf.where([[[True, False], [False, True], [True, True]]])
<tf.Tensor: shape=(4, 3), dtype=int64, numpy=
array([[0, 0, 0],
       [0, 1, 1],
       [0, 2, 0],
       [0, 2, 1]])>

Multiplexing between x and y

If x and y are provided (both have non-None values):

tf.where will choose an output shape from the shapes of condition, x, and y that all three shapes are broadcastable to.

The condition tensor acts as a mask that chooses whether the corresponding element / row in the output should be taken from x (if the element in condition is True) or y (if it is false).