Basic Examples (3) 

In NumPy, multidimensional array operations are tail-aligned:

In[1]:=

Out[1]=

Here this alignment is not supported:

In[2]:=

Out[2]=

Use Broadcasting to achieve the same effect as in NumPy:

In[3]:=

Out[3]=

Scope (3) 

Plus does not thread over arrays of different dimensions:

In[4]:=

Out[4]=

Define the Broadcasting Plus operator:

In[5]:=

Now this kind of operation can be done:

In[6]:=

Out[6]=

Properties and Relations (2) 

Listable functions in the Wolfram Language can effectively thread when initial dimensions match:

In[7]:=

Out[7]=

NumPy does not support this initial alignment:

In[8]:=

Out[8]=

Possible Issues (1) 

If dimensions are not broadcastable, then Broadcasting will report an error and will not return any results:

In[9]:=