Bookmark and Share

Mathematica is functional and has high code density. It tends to lead one to use more powerful function than the task really necessarily requires. Because high code density typically means it needs to check many things to figure out default (or Automatic) values, it slows down the program unnecessarily sometimes.

Join vs. Flatten

Use Join to join lists, not Flatten, if you know it's safe. Join is faster and uses less memory.

In[6]:= Quit[]
In[1]:= data=Table[RandomVariate[NormalDistribution[],100],{20}];
In[2]:= m1=MemoryInUse[];
Do[t=Flatten[data,1],{10000}]//AbsoluteTiming
m2=MemoryInUse[];
m2-m1
Out[3]= {3.9662268,Null}
Out[5]= 74704

Compare it to

Join@@list
In[2]:= m1=MemoryInUse[];
Do[t=Join@@data,{10000}]//AbsoluteTiming
m2=MemoryInUse[];
m2-m1
Out[3]= {0.1730099,Null}
Out[5]= 36240

blog comments powered by Disqus