[R] dplyr or plyr or both?

[Frans Marcelissen]

Hello Christopher and others
:
What cannot be stressed enough is: do not combine both packages, it gives
errors and incorrect results! I will show that below
--------------------------------------------------------
a<-data.frame(groep=1:4,v=1:40)
library(dplyr)
a %>% group_by(groep) %>% summarise(m=mean(v),n=n())
# groep     m     n
# <int> <dbl> <int>
# 1     1    19    10
# 2     2    20    10
# 3     3    21    10
# 4     4    22    10
# correct

library(plyr)
a %>% group_by(groep) %>% summarise(m=mean(v),n=n())

Error in n() : This function should not be called directly
# ???
a %>% group_by(groep) %>% summarise(m=mean(v))
# m
# 1 20.5
#incorrect!
--------------------------------------------------

So both n() and group_by from dplyr don't work after library(plyr)!

My advice is: do not use plyr. Unfortunately plyr has some functions that
are very important, and that are not in dplyr. For instance: rbind.fill()
(for combining the rows of two dataframes with unequal columns). If you
need this: do'nt library plyr, use plyr::rbind.fil

Until now I have the impression that it is also possible to library dplyr
after plyr, but it is better to remove plyr!

This is a serious problem that has been reported before, but not solved (in
dplyr 0.5.0 and plyr 1.8.4)

Frams

2016-09-15 16:09 GMT+02:00 Christopher W Ryan <cryan at binghamton.edu>:

> I've set myself the task of learning about these packages, and about
> tidy data concepts.
>
> What is the relationship between plyr and dplyr?  Does the latter
> replace the former (meaning I can concentrate on learning the latter)?
> Or is there ever a need to use functions from both (meaning I should
> learn both)?
>
> Thanks.
>
> --Chris Ryan
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

	[[alternative HTML version deleted]]