[R] Merge data under conditions

[David Winsemius]

On Mar 12, 2011, at 4:14 PM, flymer wrote:

> Dear All,
>
> Debuting in R, I'm facing a problem.
> I have 2 vectors, say 'a' et 'b', and I'd like to merge them  
> according to
> the proximity of their variable 'time'.
> How to do to keep elements which satisfy (for example) 'a$time-b 
> $time<0.5'?
>
> For example :
>
>> a
>  time x
> 1  1.0 4
> 2  2.2 5
> 3  5.2 6
>
>> b
>  time y
> 1    0 1
> 2    1 3
> 3    2 5
> 4    4 7
> 5    5 9
>
> I'd like to get :
>
>>
>  time x y
> 1  1.0 4 3
> 2  2.2 5 5
> 3  5.2 6 9
>
> I thought using the fonction 'merge'...

There are often SQL magical incantation to acheive such, and there is  
an `sqldf` package that might help, but I am not competent with it.  
Here is a base R solution using three functions (six, if you count  
"$", "<", and "-":

?expand.grid
?rep
?"["

dfrm<- expand.grid(a$time, b$time)
dfrm$x <- a$x  # by virtue of recycling
dfrm$y <- rep(b$y,  each=3)

 > dfrm[abs(dfrm$Var1-dfrm$Var2) < 0.5, ]
    Var1 Var2 x y
4   1.0    1 4 3
8   2.2    2 5 5
15  5.2    5 6 9

-- 

David Winsemius, MD
West Hartford, CT