[R] Using barplot() with zoo -- names.arg not permitted?

Thu Mar 17 14:38:26 CET 2011

I've used barplot(), including the anmes.arg parameter, on data frames
successfully, but I'm even newer to using zoo than I am to R.  :-}

I am working on a functon that accepts a data frame ("df") as its
primary argument, extracts information from it to create a zoo, then
generates a plot based on that.

The data frame has a column called "time" which is a standard UNIX time
-- the number of seconds since 01 Jan 1970 (UTC), so I treat that as a
POSIXct.  (The types of the other columns in the data frame varies
greatly, and the sequence in which they occur is not expected to be
constant.  Thus, I find it easier to let R make a reasonable guess as to
the data type, and override special cases -- such as "time" -- on an
individual basis.)

The data frame contains certain columns whose values are counters, so I
generate a zoo of the diff():

  oid = "kern.cp_time_"
  pat = paste("^", oid, sep = "")
  class(df$time) <- "POSIXct"
  df_d <- diff(zoo(df[, grep(pat, names(df))], order.by = df$time))

I then save the start & end timestamps for future reference, and
generate another zoo, this one containing only the percentages from
df_d:

  cpu_states <- c("sys", "intr", "user", "nice", "idle")
  d_range <- range(index(df_d))
  df_pct <- sweep(df_d, 1, apply(df_d, 1, sum), "/")[, paste(oid, cpu_states[1:4], sep = "")] * 100

Well, that is, I save the percentages in which I'm interested.

So far, so good.  df_pct is a zoo (as expected, and the content looks
reasonable.

I then plot it, e.g.:

barplot(df_pct, border = NA, col = colvec, space = 0, main = title,
  sub = sub, ylab = "CPU %", ylim = c(0, 100), xlab = "Time")

But the X-axis labels show up as "large integers" -- the POSIXct values
are apparently treated as numeric quantities for this purpose.

And if I try

barplot(df_pct, border = NA, col = colvec, space = 0, main = title,
  sub = sub, ylab = "CPU %", ylim = c(0, 100), xlab = "Time",
  names.arg = index(df_pct))

I am told:

Warning messages:
1: In plot.window(xlim, ylim, log = log, ...) :
  "names" is not a graphical parameter
2: In axis(if (horiz) 2 else 1, at = at.l, labels = names.arg, lty = axis.lty,  :
  "names" is not a graphical parameter
3: In title(main = main, sub = sub, xlab = xlab, ylab = ylab, ...) :
  "names" is not a graphical parameter
4: In axis(if (horiz) 1 else 2, cex.axis = cex.axis, ...) :
  "names" is not a graphical parameter

which I find a bit perplexing.

If I plot via:

barplot(df_pct, border = NA, col = col_vec, space = 0, main = title,
  sub = sub, ylab = "CPU %", ylim = c(0, 100), x lab = "Time",
  axisnames = FALSE)

that works OK, but then I'm unable to generate a human-readable set
of X-axis labels: At best, I'm able to get a single label at the origin.

Here's an excerpt from the generated df_d:
Browse[2]> df_d[1:10]
           kern.cp_time_idle kern.cp_time_intr kern.cp_time_nice
1297278017               275                 0                 0
1297278018               231                 0                 0
1297278019               266                 1                 0
1297278020               230                 2                 0
1297278021               191                 0                 0
1297278022               114                 0                 0
1297278023                30                 0                 0
1297278024                 0                 0                 0
1297278025                 0                 0                 0
1297278026                 0                 0                 0
           kern.cp_time_sys kern.cp_time_user
1297278017                8                 1
1297278018               17                35
1297278019               10                 6
1297278020               28                24
1297278021               21                72
1297278022               27               147
1297278023               43               219
1297278024               47               249
1297278025               41               259
1297278026                8                15

Here's a excerpt from the generated df_pct:

Browse[2]> df_pct[1:10]
           kern.cp_time_sys kern.cp_time_intr kern.cp_time_user
1297278017         2.816901         0.0000000         0.3521127
1297278018         6.007067         0.0000000        12.3674912
1297278019         3.533569         0.3533569         2.1201413
1297278020         9.859155         0.7042254         8.4507042
1297278021         7.394366         0.0000000        25.3521127
1297278022         9.375000         0.0000000        51.0416667
1297278023        14.726027         0.0000000        75.0000000
1297278024        15.878378         0.0000000        84.1216216
1297278025        13.666667         0.0000000        86.3333333
1297278026        34.782609         0.0000000        65.2173913
           kern.cp_time_nice
1297278017                 0
1297278018                 0
1297278019                 0
1297278020                 0
1297278021                 0
1297278022                 0
1297278023                 0
1297278024                 0
1297278025                 0
1297278026                 0

I'm using:

R version 2.12.2 (2011-02-25)

in a FreeBSD 8.2-STABLE #125 r219696: Wed Mar 16 04:19:54 PDT 2011
environment.

I'm willing to share the function (in its current state) and a sample
data set (~90KB), but didn't think it would be appropriate to spam the
list with either at this point.

Thanks!

Peace,
david
-- 
David H. Wolfskill				david at catwhisker.org
Depriving a girl or boy of an opportunity for education is evil.

See http://www.catwhisker.org/~david/publickey.gpg for my public key.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20110317/e9e697fd/attachment.bin>