[R] Problem with Aggregate - Sum, limit on number of criteria

[Uwe Ligges]

This is fixed in recent versions of R:

------------------------------------------------------------------------
r52862 | hornik | 2010-09-01 18:21:49 -0400 (Wed, 01 Sep 2010) | 1 line
Changed paths:
    M /trunk/src/library/stats/R/aggregate.R

Avoid integer overflows.



Uwe Ligges



On 22.10.2010 21:20, James Splinter wrote:
> Hello,
>
> It appears there is a limit in the number of criteria that can be put into
> the Aggregate sum function. (It looks like it is 32).
>
> My code is;
>
> HSfirst=aggregate(count,
> list(P2010W,P2009S,P2009W,P2008S,P2008W,P2007S,P2007W,P2006S,P2006W,pcom,W2010W,W2009S,W2009W,W2008S,W2008W,W2007S,W2007W,W2006S,W2006W,wcom,cd,f,g,m,urb,nourb,nourb2,nourb3,nourb4,eight,thirty,fifty,sixty,seventy,xover),sum)
> names(HSfirst)=c("P2010W","P2009S","P2009W","P2008S","P2008W","P2007S","P2007W","P2006S","P2006W","pcom","W2010W","W2009S","W2009W","W2008S","W2008W","W2007S","W2007W","W2006S","W2006W","wcom","cd","f","g","m","urb","nourb","nourb2","nourb3","nourb4","eight","thirty","fifty","sixty","seventy","xover","count")
> summary(HSfirst)
>
> some output from the summary command:
>
>       sixty      seventy      xover       count
>   Min.   :0   Min.   :0   Min.   :0   Min.   :    1.00
>   1st Qu.:0   1st Qu.:0   1st Qu.:0   1st Qu.:    1.00
>   Median :0   Median :0   Median :0   Median :    1.00
>   Mean   :0   Mean   :0   Mean   :0   Mean   :   30.35
>   3rd Qu.:0   3rd Qu.:0   3rd Qu.:0   3rd Qu.:    2.00
>   Max.   :0   Max.   :0   Max.   :0   Max.   :25016.00
>
>
> Originally I had a problem with the variable named "xover" coming out all
> zeros (a column of zeros) after the data was put through the aggregation.
> However, I verified this variable is not a column of zeros, precisely BEFORE
> it goes into the aggregation, and it all zeros afterwards. To test the
> theory, I created nourb, nourb2, nourb3 and nourb4 (which are just
> duplicates of urb) and added them, the result was that the last variables
> turned into columns of zeros.
>
> The error message is:
>   Warning messages:
> 1: In grp * nlevels(ind) : NAs produced by integer overflow
> 2: In grp * nlevels(ind) : NAs produced by integer overflow
> 3: In grp * nlevels(ind) : NAs produced by integer overflow
> 4: In grp * nlevels(ind) : NAs produced by integer overflow
>
> I tried searching for this error message but didn't find anything. I have
> very little programming background, but my colleague reviewed the
> source-code and said that he didn't find anything putting a limit on the
> number of criteria, despite that it seemingly is 32. I was hoping to find if
> this could be verified, that if there is or isn't a limit. Aswell, if there
> is a limit, what action can we take to fix the code.
>
> If anymore information is needed, please ask.
>
> Thanks,
>
> James
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.