3

I am familiar with using expression() and bquote() when plotting special Tex like markup in textual base graphics in R. But how to do "approximately equal to" without having the parser throw an error. I am referring to ?plotmath when determining the markup.

plot(c(1,100), c(1,100), type="n", axes=FALSE, xlab="", ylab="")

# This works
text(50, 80, expression(paste("The area is 24 ", km^2)))

# This doesn't
text(50, 50, expression(paste("Reginald is ", %~~% , "14 years old")))
plannapus
  • 18,529
  • 4
  • 72
  • 94
digitalmaps
  • 2,885
  • 3
  • 22
  • 28

1 Answers1

11

This should do the trick:

plot(c(1,100), c(1,100), type="n", axes=FALSE, xlab="", ylab="")
text(50, 50, 
     expression(paste("Reginald is ", phantom() %~~% phantom(), "14 years old")))

The reason your version didn't work is that the %~~% 'operator' expects to be part of an expression with something to both its left and its right. phantom() provides you with a something that is (a) invisible and (b) takes up no width --- exactly what you need.

Edit:

I learn something every day. Brian Diggs points out, in the comments, that you can produce the same effect, more simply by just leaving out the commas in the expression (in which case you don't even need a nested call to paste()). This is the better solution:

text(50, 50, expression("Reginald is " %~~% "14 years old"))
Josh O'Brien
  • 159,210
  • 26
  • 366
  • 455
  • Thanks for this, it works. Is this documented somewhere obvious that I missed? – digitalmaps Mar 09 '12 at 21:33
  • No, it's not obvious, and it was hard for me to find when I first needed it. I think I discovered it by looking through the examples displayed by typing `demo(plotmath)`. It's on the third page that pops up. It's also included in the list of features `?plotmath`, but its utility isn't really spelled out. – Josh O'Brien Mar 09 '12 at 21:39
  • 4
    In this case, you can use the strings themselves instead of `phantom()`: `text(50, 50, expression(paste("Reginald is " %~~% "14 years old")))` – Brian Diggs Mar 09 '12 at 21:48
  • @BrianDiggs -- Excellent. Thanks. I've added an edited version of your suggestion to the main post. For whatever reason, your version also helps me to really 'get' for the first time the kind of thing that R must be doing to produce the mathematical annotation it adds to plots. So thanks for that as well. – Josh O'Brien Mar 09 '12 at 21:57