column name missing in process_stats

[HarlanH]

Hey Hadley. I'm playing around with gg2v -- cool stuff. Actually planning to do some stuff with Vega in another language for work, but using gg2v to do some quick and dirty proof of concept work. I've found that some simple plots that work fine when rendered via ggplot break in gg2v, seemingly because column names aren't correctly set in aggregation. Here's an example:

> library(plyr)
> library(ggplot2)
> library(gg2v)
> data(baseball)
> p <- ggplot(baseball) + layer(mapping=aes_string(x='lg',y='r'), geom='bar', stat='summary', fun.y='sum')
> p                                 # this works
> gg2v(plot=p)                 # this doesn't
Error in eval(expr, envir, enclos) : object 'y' not found
> options(error=recover)
> gg2v(plot=p)
Error in eval(expr, envir, enclos) : object 'y' not found

Enter a frame number, or 0 to exit   

1: gg2v(plot = p)
2: plot_spec(plot, data_dir = data_dir, ...)
3: plot_data(plot)
4: Map(function(data, aes) render_data(data, aes, env = plot$plot_env), udata,
5: mapply(FUN = f, ..., SIMPLIFY = FALSE)
6: (function (data, aes) 
render_data(data, aes, env = plot$plot_env))(dots[[1]
7: render_data(data, aes, env = plot$plot_env)
8: eval(var, data, env)
9: eval(expr, envir, enclos)

Selection: 3
Called from: eval(var, data, env)
Browse[1]> head(data)
[[1]]
   x group     V1
1        1   2745
2 AA     2   9563
3 AL     3 299992
4 FL     4   1025
5 NL     5 373241
6 PL     6   2739
7 UA     7    300

Browse[1]> head(aes)
$f5f5faf07687c3b68ede8a16ee7e0bce
$f5f5faf07687c3b68ede8a16ee7e0bce$x
x

$f5f5faf07687c3b68ede8a16ee7e0bce$y
y

I'd like to get this fixed for my demos. The issue seems to be in plot-data.R::process_stats, but I'm not exactly sure what needs to be done. Any suggestions? Thanks!

-Harlan

[HarlanH]

FWIW, using this function instead of the convert in your utils.R seems to get around this problem, although I have no idea if it's the preferred solution:

combine2 <- function(x, y) {
  n <- length(x)
  for(i in seq_along(y)) {
    x[[n + i]] <- y[[i]]
    names(x)[[n+i]] <- names(y)[[i]]
  }
  x
}

[hadley]

Hi Harlan,

gg2v is basically just a proof of concept that you could convert ggplot2 into vega, but it's not something I'll be developing further. You're more than welcome to fork and continue working on it though.

Hadley