Filtering on a column with NaNs multiple times returns variable numbers of rows

[tleonardi]

test <- data.frame( Name = c("a", "b","c", "d", "e"),
                    ID = c(1, 1, 1, 1, 1),
                    expression = c(NaN, NaN, NaN, NaN, NaN)
                  )

rep <- replicate( 5000, nrow(group_by(test, ID) %>% filter(min_rank(desc(expression))==1)))

table(rep)

>rep
>   1    2    5
>4654  345    1

sessionInfo()

R version 3.1.0 (2014-04-10)
Platform: x86_64-unknown-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C               LC_TIME=en_GB.UTF-8        LC_COLLATE=en_GB.UTF-8     LC_MONETARY=en_GB.UTF-8    LC_MESSAGES=en_GB.UTF-8    LC_PAPER=en_GB.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] dplyr_0.3.0.2

loaded via a namespace (and not attached):
[1] assertthat_0.1 DBI_0.3.1      lazyeval_0.1.9 magrittr_1.0.1 parallel_3.1.0 Rcpp_0.11.3    tools_3.1.0

[romainfrancois]

Hum. Another way to exhibit the problem:

> replicate( 10, { data <- group_by(test, ID) %>% mutate(rank = min_rank(expression) ); data$rank  } )
           [,1]       [,2]       [,3]       [,4]       [,5]       [,6]
[1,]          1          1          1          1          1          1
[2,]      32735      32735      32735      32735      32735      32735
[3,] 1384250248 1384250248 1384250248 1384250248 1384250248 1381406080
[4,]      32735      32735      32735      32735      32735      32735
[5,] 1384169064 1384169064 1384169064 1384169064 1384169064 1381409544
           [,7]       [,8] [,9] [,10]
[1,]          1          1    1     1
[2,]      32735 1072693248    3     1
[3,] 1381406120          0  100     1
[4,]      32735 1074790400   39     1
[5,] 1381409584          0    3     1

In this example, only the 10th result is what I'd expect :/