.. meta::
    :description: Code snippets Numpy Array create channel

How to create a Channel from a Numpy Array
==========================================

In this example, a Channel is created from a Numpy array.

Importing the Numpy package and creating the array::

    import numpy as np
    # Creating the data matrix: initialization with a specific data type
    newNumpyArrayMatrix = np.zeros([4, 5, 6], dtype='int32')
    # Making a cross in the data
    newNumpyArrayMatrix[2, :, 3] = 5
    newNumpyArrayMatrix[2, 2, :] = 5

Creating the Channel from the array::

    from ORSModel import createChannelFromNumpyArray
    newChannel = createChannelFromNumpyArray(newNumpyArrayMatrix)

.. warning::

    The method *createChannelFromNumpyArray* accepts only these Numpy data types:
    *np.int8*, *np.uint8*, *np.int16*, *np.uint16*, *np.int32*, *np.uint32* and *np.float32*.

    Note that the Numpy method *astype* can be used to convert the data type of a matrix, for example::

        aFloat64Matrix = np.zeros([4, 5, 6])
        aFloat32Matrix = aFloat64Matrix.astype(np.float32)

Setting optional properties::

    newChannel.setTitle('My dataset')

Using a *publish* to make the application aware of this new dataset::

    newChannel.publish()