Meld

The meld operation merges the hypercubes of its input tensors. The value at a given hypercube value is given by the value of the first tensor that has a value defined at that hypercube point. So ordering of input tensors does matter where the data is inconsistent between input tensors.

For example, consider the following inputs and x-vectors:

1. $\{1.0, 2.0, 3.0\}$ with x-vector $\{1,2,3\}$, and
2. $\{0.0,1.5,4.0\}$ with x-vector $\{0,2,4\}$

then the resultant output has x-vector $\{0,1,2,3,4\}$ and the values are $\{0.0,1.0,2.0,3.0,4.0\}$ if 1 is connected to port 1 and 2 connected to port 2. If they were connected the other way around, the the values would be $\{0.0,1.0,1.5,3.0,4.0\}$.