Please try 1.19.20. This is a pretty big change, so hopefully it plays well with everyone's code. There's some chance that duplicate row names can be introduced (and throw an error) while before they were simply dropped. This is because we take the names from any vector argument as potential rownames, just like data.frame. On Thu, Aug 1, 2013 at 1:07 PM, Kasper Daniel Hansen < kasperdanielhansen@gmail.com> wrote: > Well, I was bitten by this by some custom code for manipulating the > colData slot of a SummarizedExperiment - for this class, the sampleNames > are stored exactly as the rownames of @colData so it becomes pretty > important to keep. Now that I know, it is easy to work around in the code > I had - I just save the rownames. > > Still, I was very surprised by this, and I think we should keep the > rownames. > > Best, > Kasper > > > On Thu, Aug 1, 2013 at 3:58 PM, Michael Lawrence < > lawrence.michael@gene.com> wrote: > >> In general, DataFrame does less with rownames compared to data.frame. >> This was for simplicity and performance. So we could add that for the sake >> of consistency; it was just ignored at the beginning. >> >> >> On Thu, Aug 1, 2013 at 12:35 PM, Kasper Daniel Hansen < >> kasperdanielhansen@gmail.com> wrote: >> >>> > df1 = DataFrame(A = c(1,2)) >>> > df2 = DataFrame(B = c(1,2)) >>> > rownames(df1) = c("a", "b") >>> > df1 >>> DataFrame with 2 rows and 1 column >>> A >>> <numeric> >>> a 1 >>> b 2 >>> > cbind(df1, df2) >>> DataFrame with 2 rows and 2 columns >>> A B >>> <numeric> <numeric> >>> 1 1 1 >>> 2 2 2 >>> >>> rownames are removed. This does not happen for data.frame's. >>> >>> Kasper >>> >>> [[alternative HTML version deleted]] >>> >>> _______________________________________________ >>> Bioconductor mailing list >>> Bioconductor@r-project.org >>> https://stat.ethz.ch/mailman/listinfo/bioconductor >>> Search the archives: >>> http://news.gmane.org/gmane.science.biology.informatics.conductor >>> >> >> > [[alternative HTML version deleted]]