Question

I had asked this question earlier and thought I might post an example that shows why I was seeing the effect, in the event that it might be helpful:

require(data.table)
x <- data.table(a=1:10)
x[,a:=scale(a)]
#    [,1]      
# 1: -1.4863011
# 2: -1.1560120
# 3: -0.8257228
# 4: -0.4954337
# 5: -0.1651446
# 6:  0.1651446
# 7:  0.4954337
# 8:  0.8257228
# 9:  1.1560120
#10:  1.4863011
na.omit(x)
Error in `[.data.table`(object, !omit) : 
  i is invalid type (matrix). Perhaps in future a 2 column matrix could return a list     of elements of DT (in the spirit of A[B] in FAQ 2.14). Please let datatable-help know if     you'd like this, or add your comments to FR #1611.

The reason seems to be that scale does not return a vector and data.table did not complain. Doing x[,a:=as.vector(scale(a))] instead appears to fix the issue. Have I missed something in the documentation?

Answer 1

Good question (I edited slightly). It should either be a warning, or data.table could coerce the 1 column matrix to a vector automatically, since I guess scale in the way you showed is a common task and natural to do like that. The na.omit aspect is one way of revealing the issue, but there are probably other ways since the root cause is further up as you nicely showed.

Bug report filed, thanks :

Bug#2333 := is able to create a “matrix” column, but “matrix” columns are invalid

---

Update: Root cause is now fixed in v1.8.3. A 1 column matrix is silently treated as vector. A matrix with 2 or more columns gives a warning.