Re: multiple outputs from a function

*To*: mathgroup at smc.vnet.net*Subject*: [mg52768] Re: multiple outputs from a function*From*: David Bailey <dave at Remove_Thisdbailey.co.uk>*Date*: Mon, 13 Dec 2004 04:22:04 -0500 (EST)*References*: <cpekeh$744$1@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

Ben Barrowes wrote: > I feel I must be missing something fundamental... > > How do you write a function (or ?) in Mathematica which produces more > than one output? > > Let's say I have some data and I want a single function to calculate the > mean, variance, std, etc, more than one thing and return those? I know > that there are builtin functions for those parameters, but the point is > I want to define functions with more than one output. > > The only examples I can find are along the lines of: > > f[x_,y_]:=x^2+y^2; > > which have only a single result. > > Is there a different structure altogether, such as a Subroutine, which > allows multiple results from a single subunit? > > One thought I had was that because Mathematica treats everything as > global unless defined specifically local (e.g. in a module), that > variables used in a procedure would be accessible and would thus be a > "result", but it seems scoping problems would arise if this was used too > often. For example: > > In[67]:= > t1[x_,y_]:=(a1=x^2+y;a2=x+y^2;x+y) > > In[68]:= > t2=t1[5,6] > > Out[68]= > 11 > > In[69]:= > a1 > > Out[69]= > 31 > > In[70]:= > a2 > > Out[70]= > 41 > > Is this the accepted method for extracting multiple results from one > function definition? > > > Ben Barrowes > Hi, It is sometimes useful to arrange for data to come back in global variables, as you have done - although it would be best to use more distinctive names to avoid confusion. In most situations it is best to return a list of items, or a structure - using a head with no definition. Thus, in your case you might return something like myStatics[5.2,1.1] where the first number is the mean and the other is the standard deviation. Obviously, you can extend this to return as much information as you like. Regards, David Bailey dbaileyconsultancy.co.uk