dplyr: cannot subset within summarise?

[Jonathan Dobres]

Hi there!

I'm a big plyr fan who's trying to make the switch to dplyr, but I've run into a deal-breaker issue. plyr has the ability to subset within summarise(), but it looks like dplyr doesn't. For example:

diamond.plyr <- ddply(diamonds, .(cut), summarise, 

                      all=mean(price), 

                      E=mean(price[color == 'E']));

Produces:

        cut      all        E

1      Fair 4358.758 3682.312

2      Good 3928.864 3423.644

3 Very Good 3981.760 3214.652

4   Premium 4584.258 3538.914

5     Ideal 3457.542 2597.550

But, the equivalent with dplyr:

diamond.dplyr <- diamonds %.%

    group_by(cut) %.%

    dplyr::summarise(all=mean(price),

              E=mean(price[color == 'E']));

Gives:

        cut      all   E

1      Fair 4358.758 NaN

2 Very Good 3981.760 NaN

3      Good 3928.864 NaN

4   Premium 4584.258 NaN

5     Ideal 3457.542 NaN

Am I misunderstanding something about dplyr's syntax? Is this type of internal subsetting no longer valid, or is there another way to do it? It's a succinct shortcut that's saved me a lot of time and kept my code very readable, so despite dplyr's obvious speed improvements, I can't really switch until I figure out an equivalent.

-Jon

[Hadley Wickham]

Hi Jonathan,

It works fine for me:

library(dplyr)

data(diamonds, package = "ggplot2")

diamonds %.%

  group_by(cut) %.%

  summarise(all=mean(price), E=mean(price[color == 'E']));

Source: local data frame [5 x 3]

        cut      all        E

1      Fair 4358.758 3682.312

2 Very Good 3981.760 3214.652

3      Good 3928.864 3423.644

4   Premium 4584.258 3538.914

5     Ideal 3457.542 2597.550

Maybe some plyr/dplyr interference?

Hadley

--
You received this message because you are subscribed to the Google Groups "manipulatr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to manipulatr+...@googlegroups.com.
To post to this group, send email to manip...@googlegroups.com.
Visit this group at http://groups.google.com/group/manipulatr.
For more options, visit https://groups.google.com/groups/opt_out.

--
http://had.co.nz/

[Jonathan Dobres]

I've unloaded all the extra libraries I usually load, cleared my workspace, and copied/pasted your code directly, and I still get a column of NaNs where numbers should be. At this point I have no idea what's going on. I'm also running the latest versions of R and RStudio for Mac, on Mac OS 10.7.5.

[Winston Chang]

I get the same NaN when I use the CRAN version of dplyr. With the development version from Github, I get the correct numeric values.

-Winston

On Thu, Jan 23, 2014 at 8:18 PM, Jonathan Dobres <jdo...@gmail.com> wrote:

I've unloaded all the extra libraries I usually load, cleared my workspace, and copied/pasted your code directly, and I still get a column of NaNs where numbers should be. At this point I have no idea what's going on. I'm also running the latest versions of R and RStudio for Mac, on Mac OS 10.7.5.

--

[Jonathan Dobres]

Interesting. I had been using the CRAN version. I'll try the Github version tomorrow.

[Jonathan Dobres]

Grabbing the version on Github does seem to have resolved the problem. Looks like the issue is specific to the version on CRAN.