Question

I have a data frame like

    A   B   C   D   E   F
2003-07 445 48  1126    512 277 677
2003-08 429 77  1179    583 320 675
2003-09 421 80  1077    488 288 627

I want to plot these time series on the same graph. I also want A B C to have the same colour (blue) and the others to be black. So I use

ts.plot(df,col=c(rep("blue",3),rep("black",3)))

This creates the correct time series plot, except the colours are applied to the wrong series: A B C are a mix of blue and black and so are D E F. (Note that the real data frame has a lot more columns if that is the source of the problem.)

What order does ts.plot() use?

EDIT

df <- structure(list(X = structure(1:3, 
.Label = c("2003-07", "2003-08", "2003-09"), class = "factor"), 
 A = c(445L, 429L, 421L), B = c(48L, 77L, 80L),
 C = c(1126L, 1179L, 1077L), D = c(512L, 583L, 488L ), 
 E = c(277L, 320L, 288L), FF = c(677L, 675L, 627L)), 
.Names = c("X", "A", "B", "C", "D", "E", "FF"), 
class = "data.frame", row.names = c(NA, -3L))

Answer 1

ts.plot(df,col=c(rep("blue",3),rep("black",3)))

I think it is plotting absolutely as specified in the order A-B-C-D-E-F.
So…

A     B     C     D     E     F
Blue  Blue  Blue  Black Black Black 

If you match up your values for A/B/C…

          A  B    C
2003-07 445 48 1126
2003-08 429 77 1179
2003-09 421 80 1077

…you will see the three blue lines are those going from 445->429->421 & 48->77->80 & 1126->1179->1077, which actually matches what you specified you want.

As an aside, you could simplify your plot call to be:

ts.plot(df,col=rep(c("black","blue"),each=3))

…by using the each= part of the function call.

EDIT::

Looking at your dput, it looks like your graph is also plotting your labels df$X as the values 1,2,3 since it is a factor and gets interpreted as such when plotting.

Try this bit of code to remove the labels from your original plot and add them back in again:

ts.plot(df[-1],col=c(rep("blue",3),rep("black",3)),gpars=list(xaxt="n"))
axis(1,labels=as.character(df$X),at=1:3)