[R] Selecting closest values

[Frank E Harrell Jr]

Bill.Venables at csiro.au wrote:
> Here is one way you might find useful
> 
>> id <- 1:183
>> lo <- sample((-22):144, 183, rep = TRUE)
>> ds <- outer(lo, lo, function(x,y) abs(x-y))
>> rs <- t(apply(ds, 2, function(x) id[order(x)][2:11]))
>> rownames(rs) <- id
> 
> Here's what I got
> 
>> head(rs)
>   [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
> 1  118   23  125  182  183    8   86  119  141   175
> 2   27  167  176    4  126   45  113   81  144   153
> 3   43    7   18   60  148   11   12  150   85   163
> 4  126   27  167  176    2   45  144  153   48    55
> 5   15   46   86  141  175   23  125  183    1   118
> 6   58   78  139   13   66  131  156  181  117   157
> 
> If you want anything more random in the case of ties than this, you will have to work a bit harder on the code.  
> 
> W.
> 
> Bill Venables
> http://www.cmis.csiro.au/bill.venables/ 
> 
> 
> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of P_M
> Sent: Monday, 23 March 2009 1:32 AM
> To: r-help at r-project.org
> Subject: [R] Selecting closest values
> 
> 
> Hi
> I have a table with ID (1 to 183) and Location (144 to -22).
> My problem is that I want to select the 10 ID's that are closest in Location
> to ID 1, ID 2 and so on.
> Also, some ID have the same Location. Say, if 11 ID's are closest to ID 100
> I want to randomly choose one of the ID's to select 10 ID's total.
> 
> Thank you
> 
>  
> 

Also take a look at the whichClosest function in the Hmisc package.  But 
whichClosest cheats by calling Fortran.

Cheers
Frank

-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University