[Rd] reversing xlim, ylim in smoothScatter

Martin Maechler maechler at stat.math.ethz.ch
Wed Apr 22 19:29:05 CEST 2009

>>>>> "SG" == Shashikiran Ganesh <shashikiran_ganesh at yahoo.co.in>
>>>>>     on Wed, 22 Apr 2009 17:16:09 +0530 (IST) writes:

    SG> Hello,
    SG> I have found that in smoothScatter it is not possible to reverse the axes plotted (R version 2.9.0) .   It appears that this arises from the hard coding of xlim and ylim in smoothscatter.R in the lines :

    SG> x <- x[ xlim[1] <= x[,1] & x[,1] <=xlim[2], ]   (line  number 25)

    SG> and 

    SG> x <- x[ ylim[1] <= x[,2] & x[,2] <= ylim[2], ]  (line number 31)

    SG> This results in a x being NA if ylim[1] > ylim[2] which results in an error on executing
    SG>  map <- grDevices:::.smoothScatterCalcDensity(x, nbin, bandwidth)

    SG> To counter this problem, I replaced the above two lines by :
    SG> x <-x [min(xlim) <= x[,1] & x[,1] <= max(xlim), ]

    SG> and 
    SG> x <- x[min(ylim) <= x[,2] & x[,2] <= max(ylim), ]

    SG> and now smoothscatter reverses axes properly if xlim and/or ylim are provided with reversed ranges.   

    SG> I am wondering if this would break something somewhere else or if there is a better way to reverse the axes.     (I am a newbie in R programming although I have been using R for some astronomical plots and computations etc...) 

No, I don't think your proposed change could create problems,
but to the contrary, I'd declare it as a simple bug fix,
so it will be part of the next versions of R,
thank you very much!

Martin Maechler, ETH Zurich

    SG> Thanks in advance for your comments!  
    SG> Shashi
    SG> --
    SG> Shashikiran Ganesh 

    SG> http://www.prl.res.in/~shashi 

    SG> http://cosmicdiary.org/blogs/shashikiran_ganesh/
    SG> Registered linux user number: 39542 (http://counter.li.org)

More information about the R-devel mailing list