Stacked barplot of timeseries data

On Tue, 3 Jun 2008, Demetri S. Mouratis wrote:

Hi,

I'm trying to plot time-series data where each sample breaks down the
percentage of CPU time spent in each of four states (usr, nice, sys, idle)


19:08:15  %usr  %nice   %sys   %idle
19:08:16    5      0     10     86
19:08:17   17      0     14     69
19:08:18    5      0      8     87
19:08:19   10      0     10     81
19:08:20    3      0      7     90
19:08:21    4      0      8     88
[on and on for many samples]

The plot I'm aiming for would stack the first three states in a colored
barplot, so you get a visual sense of how busy the system is over the course
of the day, and which state the CPU is spending its time in.  (I've done
this as area charts as well).

barplot() looked promising, but it wants to stack the columns instead of
the rows.

Anybody have a good solution for this?

The "zoo" package provides a barplot() method for "zoo" series.

Example with some artificial data:

 library("zoo")
 x <- matrix(runif(44), ncol = 4)
 colnames(x) <- c("%usr", "%nice", "%sys", "%idle")
 z <- zoo(x, Sys.time() - c(10:0))
 barplot(z, legend = TRUE)

hth,
Z

Thanks!
...Demetri