Multidimensional python arrays

If arrays of array.array or bytearray , then nothing. Python does not have multidimensional arrays.

Unless to make a one-dimensional array of size n*n , and to address on an index x+n*y . For example, like bytearray(n*n) or, say, array.array("l", [0 for _ in range(0, n*n)]) .

If the concept of a “list of lists” fits the concept of a “multidimensional array” (lists in CPython are implemented exactly as arrays ), then, in fact, make them. For example:

 >>> n = 3 >>> [[0 for _ in range(0, n)] for _ in range(0, n)] [[0, 0, 0], [0, 0, 0], [0, 0, 0]] 

If matrices are needed, then NumPy can be used:

 >>> numpy.matrix(numpy.zeros((n, n))) matrix([[ 0., 0., 0.], [ 0., 0., 0.], [ 0., 0., 0.]]) 

Or, if not matrices are needed (and n × n was a special case), then directly, numpy.array :

 >>> numpy.zeros((n, n)) array([[ 0., 0., 0.], [ 0., 0., 0.], [ 0., 0., 0.]]) 

Or, as @mikillskegg correctly notes, for example, if an array of non-numbers is required, instead of numpy.array / numpy.zeros you can use the lower-level interface numpy.ndarray :

 >>> a = numpy.ndarray(shape=(n, n), dtype=(unicode, 1)) >>> a.fill(u"X") array([[u'X', u'X', u'X'], [u'X', u'X', u'X'], [u'X', u'X', u'X']], dtype='<U1') 

In general, the question is strangely formulated, because neither the word is said about the data that should be stored in the structure, nor about the properties that the structure should have.