This cannot be done unless transitivity is guaranteed. Is it? S L
a b b c c a Bert On Thu, Mar 14, 2019, 4:30 AM Pedro Conte de Barros <pbar...@ualg.pt> wrote: > Dear All, > > This should be a quite established algorithm, but I have been searching > for a couple days already without finding any satisfactory solution. > > I have a matrix defining pairs of Smaller-Larger arbitrary character > values, like below > > Smaller <- c("ASD", "DFE", "ASD", "SDR", "EDF", "ASD") > > Larger <- c("SDR", "EDF", "KLM", "KLM", "SDR", "EDF" > > matComp <- cbind(Smaller, Larger) > > so that matComp looks like this > > Smaller Larger > [1,] "ASD" "SDR" > [2,] "DFE" "EDF" > [3,] "ASD" "KLM" > [4,] "SDR" "KLM" > [5,] "EDF" "SDR" > [6,] "ASD" "EDF" > > This matrix establishes six pairs of "larger than" relationships that > can be used to sort the unique values in the matrix, > > > unique(as.vector(matComp)) > [1] "ASD" "DFE" "SDR" "EDF" "KLM" > > Specifically, I would like to get this: > > sorted <- c("ASD", "DFE", "EDF", "SDR", "KLM") > > or, equally valid (my matrix does not have the full information): > > sorted <- c("DFE", "ASD", "EDF", "SDR", "KLM") > > Preferably, I would get the different combinations of the unique values > that satisfy the "larger than" conditions in the matrix... > > > I am sure this is a trivial problem, but I could not find any algorithm > to solve it. > > Any help would be highly appreciated > > ______________________________________________ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > 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. > [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.