Method int

python/dgl/heterograph.py:6351–6399

Cast the graph to one with idtype int32 If the graph already has idtype int32, the function directly returns it. Otherwise, it returns a cloned graph of idtype int32 with features copied (shallow copy). Returns ------- DGLGraph The graph of idtyp

(self)

        return self.astype(F.int64)

    def int(self):
        """Cast the graph to one with idtype int32

        If the graph already has idtype int32, the function directly returns it. Otherwise,
        it returns a cloned graph of idtype int32 with features copied (shallow copy).

        Returns
        -------
        DGLGraph
            The graph of idtype int32.

        Examples
        --------

        The following example uses PyTorch backend.

        >>> import dgl
        >>> import torch

        Create a graph of idtype int64.

        >>> # (0, 1), (0, 2), (1, 2)
        >>> g = dgl.graph((torch.tensor([0, 0, 1]), torch.tensor([1, 2, 2])))
        >>> g.ndata['feat'] = torch.ones(3, 1)
        >>> g.idtype
        torch.int64

        Cast the graph to one of idtype int32.

        >>> # A cloned graph with an idtype of int32
        >>> g_int = g.int()
        >>> g_int.idtype
        torch.int32
        >>> # The idtype of the original graph does not change.
        >>> g.idtype
        torch.int64
        >>> g_int.edges()
        (tensor([0, 0, 1], dtype=torch.int32), tensor([1, 2, 2], dtype=torch.int32))
        >>> g_int.ndata
        {'feat': tensor([[1.],
                         [1.],
                         [1.]])}

        See Also
        --------
        long
        idtype
        """
        return self.astype(F.int32)


############################################################