[R] Plot two values on same axes

[Rolf Turner]

On 27/09/19 10:27 AM, Rich Shepard wrote:

> I want to plot maximum and minimum water temperatures on the same axes and
> thought I had the correct syntax:
> 
> watertemp <- read.table("../../data/hydro/water-temp.dat", header = 
> TRUE, sep =",")
> watertemp$sampdate <- as.Date(as.character(watertemp$sampdate))
> watertempsum <- summary(watertemp)
> print(watertempsum)
> maxwatemp <- xyplot(maxtemp ~ sampdate, data=watertemp, col="red", 
> type="h", main="USGS Foss Gauge Water Temperatures, 1974-1981", 
> ylab="Temperature (C)", xlab="Date", scales=list(tck=c(1,0)))
> minwatemp <- xyplot(mintemp ~ sampdate, data=watertemp, col="blue", 
> type="h")
> plot(maxwatemp)
> par(new=TRUE)
> plot(minwatemp)
> 
> However, R plots only minwatemp and displays this:
> Warning message:
> In par(new = TRUE) : calling par(new=TRUE) with no plot
> 
> Despite my searching for the reason I don't see what syntax error I made.
> The two questions I ask of you:
> 
> 1. Where have I gone wrong in this script?
> 
> 2. With 2,492 daily observations in the source file (including 242 NAs for
> maxtemp and 243 NAs for mintemp), what would be a more appropriate plot to
> show both sets of data?

You are using xyplot() from the lattice package (which you did not 
mention).  Your par(new=TRUE) syntax would probably work with base R 
graphics (although it would be suboptimal; one would normally use 
points() to add a second graph on the same figure).  I would hazard that 
it won't work at all with lattice graphics.

Moreover you seem to be trying to mix lattice graphics and base R 
graphics in a higgledy-piggledy fashion:  first you call xyplot() and 
then you call plot().

To do what you want with lattice graphics I think you need to learn 
about and apply *panels* --- in particular panel.points().  Using panels 
appropriately is a bit tricky, I find, and requires some study.

You may better off using base R graphics, unless there are 
considerations which demand the use of lattice.  Something like this, maybe:

plot(maxtemp ~ sampdate, data=watertemp, type="h", col="red",
                          <annotation arguments>)
points(maxtemp ~ sampdate, data=watertemp, type="h",col="blue")

(Not tested since your example is not reproducible.)

cheers,

Rolf Turner

-- 
Honorary Research Fellow
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276