[R] summarized data set - how to use an "occurs" field

[David Winsemius]

On Jul 15, 2011, at 6:10 PM, mloxton wrote:

> I have a data set with 22 fields and several thousand records in which
> one field (count) indicates the number of times that each specific
> combination of the other 21 fields occurred in a bigger and largely
> unavailable data set.
> So each record is unique in its combination of field values and has a
> field that identifies how many multiples of this record actually
> occurred.
>
> Without resorting to writing a program that re-expands the data set to
> several million rows by cloning each row by the number of times the
> "count" field indicated, is there a way in R to use that field to come
> up with summary stats and bargraphs of the distribution of any one of
> the other fields?
>

 > dfrm <- expand.grid(A=1:3, B=1:3)
 > dfrm$counts <- 1:9
 > xtabs(counts~A, data=dfrm)
A
  1  2  3
12 15 18

 >barplot(xtabs(counts~A, data=dfrm), xlab="Counts by A level")

-- 

David Winsemius, MD
West Hartford, CT