R help - Aug 2021 - Find "undirected" duplicates in a tibble

Eric,
> x %>% transmute( a=pmin(Source,Target), b=pmax(Source,Target)) %>%
>   unique() %>% rename(Source=a, Target=b)
ah, very nice.  i have trouble remembering, e.g., unique().

fwiw, (hopefully) here's a baser version.
----
  x = data.frame(Source=rep(1:3,4),
Target=c(rep(1,3),rep(2,3),rep(3,3),rep(4,3)))

  y <- apply(x, 1, function(y) return (c(A=min(y), B=max(y))))
  unique(t(y))
----

cheers, Greg

Hello,

This seems elegant to me but it's also the slowest, courtesy sort.

apply(x, 1, sort) |> t() |> unique()


(My tests show that for small inputs Greg's base apply is fastest, with 
nrow(x) > 700, Eric's dplyr is fastest)


Hope this helps,

Rui Barradas

?s 15:13 de 20/08/21, Greg Minshall escreveu:
> Eric,
> 
>> x %>% transmute( a=pmin(Source,Target), b=pmax(Source,Target))
%>%
>>    unique() %>% rename(Source=a, Target=b)
> 
> ah, very nice.  i have trouble remembering, e.g., unique().
> 
> fwiw, (hopefully) here's a baser version.
> ----
>    x = data.frame(Source=rep(1:3,4),
Target=c(rep(1,3),rep(2,3),rep(3,3),rep(4,3)))
> 
>    y <- apply(x, 1, function(y) return (c(A=min(y), B=max(y))))
>    unique(t(y))
> ----
> 
> cheers, Greg
> 
> ______________________________________________
> 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.
>

Note that:

1. Your solution returns a matrix, not a data frame (or Tibble)
2. Assuming that the order of the entries in the pairs does not matter
(which your solution also assumes and seems reasonable given the OP's
specification), I think that you'll find the following, which returns
the data.frame, is considerably more efficient:

x[!duplicated(cbind(do.call(pmin,x), do.call(pmax,x))),]

For example:
> x <- expand.grid(Source = 1:1000, Target = 1:1000)
>  system.time({
 y <- apply(x, 1, function(y) return (c(A=min(y), B=max(y))))
 unique(t(y))})
   user  system elapsed
  5.075   0.034   5.109
> system.time({
 x[!duplicated(cbind(do.call(pmin, x), do.call(pmax, x))), ]
 })
   user  system elapsed
  1.340   0.013   1.353

Still more efficient and still returning a data frame is:

w <- x[,2] > x[,1]
x[w,] <- x[w, 2:1]
unique(x)
> system.time({
 w <- x[, 2] > x[,1]
 x[w, ] <- x[w, 2:1]
 unique(x)})
   user  system elapsed
  0.693   0.011   0.703


The efficiency gains are due to vectorization and the use of more
efficient primitives. None of this may matter of course, but it seemed
worth mentioning.

Cheers,
Bert

Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

On Fri, Aug 20, 2021 at 7:13 AM Greg Minshall <minshall at umich.edu>
wrote:
>
> Eric,
>
> > x %>% transmute( a=pmin(Source,Target), b=pmax(Source,Target))
%>%
> >   unique() %>% rename(Source=a, Target=b)
>
> ah, very nice.  i have trouble remembering, e.g., unique().
>
> fwiw, (hopefully) here's a baser version.
> ----
>   x = data.frame(Source=rep(1:3,4),
Target=c(rep(1,3),rep(2,3),rep(3,3),rep(4,3)))
>
>   y <- apply(x, 1, function(y) return (c(A=min(y), B=max(y))))
>   unique(t(y))
> ----
>
> cheers, Greg
>
> ______________________________________________
> 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.