[datatable-help] Another CJ anomaly

[Damian Betebenner]

Not sure if this is corrected in the latest build on R-Forge but will report it anyway:

CJ doesn't do the cross product correctly:

require(data.table)

Y <- factor(1:2, levels=1:2, labels=c("MATHEMATICS", "READING"))
Z <- factor(1:3, levels=1:3, labels=paste(2008:2010, 2009:2011, sep="_"))

tmp.cj <- CJ(Y,Z)

> tmp.cj
              V1        V2
[1,] MATHEMATICS 2008_2009
[2,] MATHEMATICS 2009_2010
[3,] MATHEMATICS 2010_2011
[4,]     READING 2008_2009
[5,]     READING 2009_2010
[6,]     READING 2010_2011


### But subsetting tmp.cj shows weird behavior (the V2 levels disappear in some cases)

> tmp.cj[1:3]
              V1        V2
[1,] MATHEMATICS 2008_2009
[2,] MATHEMATICS 2009_2010
[3,] MATHEMATICS 2010_2011


> tmp.cj[4:6]
          V1   V2
[1,] READING <NA>
[2,] READING <NA>
[3,] READING <NA>


Any insight on the above greatly appreciated

Damian


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/datatable-help/attachments/20111204/b3520be5/attachment.htm>