r - Using summarise() to count the number of times the min value is repeated

Question

I have this reach data frame with ordered values and Reachability and my desired output is a summary table of several properties grouped by Cluster. The entire table contains more values but I think 10 rows are more than enough to explain what I want to achieve.

# A tibble: 500 x 3
  Order Reachability Cluster
   <int>        <dbl>   <dbl>
 1     1       NA           1
 2     2        1.54        1
 3     3        1.54        1
 4     4        0.860       1
 5     5        0.821       1
 6     6        0.821       1
 7     7        0.821       1
 8     8        0.821       1
 9     9        0.821       1
10    10        0.821       1
# ... with 490 more rows

I create my summary table with some position information about my reach table.

reach %>% dplyr::group_by(Cluster) %>% 
    summarise(first_value = first(na.omit(Reachability)),
              min_value = min(na.omit(Reachability)),
              last_value = last(na.omit(Reachability)),
              first_pos = first(Order),
              min_pos = Order[which.min(Reachability)],
              last_pos = last(Order))

# A tibble: 1 x 7
  Cluster first_value min_value last_value first_pos min_pos last_pos
    <dbl>       <dbl>     <dbl>      <dbl>     <int>   <int>   <int>
1       1       1.54      0.821      0.821       1       5      10

What I'm having trouble with is a command inside summarise that allows me to count the number of times that "min_value" repeats. In this case, for 0.821 the "min_value" should be 6. This is what I've tried with no success:

... %>% 
summarise(...
          ...
          N_min = sum(Reachability == min(na.omit(Reachability))))

... %>% 
summarise(...
          ...
          N_min = count(min(na.omit(Reachability))))

Am I missing something? I really have no idea why does my first option not work. From what I understand if I make that sum, performed by groups, should give me a sum of TRUE's (or 1's) that meet my condition. Thanks!

Data:

reach <- structure(list(Order = 1:10, Reachability = c(NA, 1.53995982068778, 
1.53995982068778, 0.860332791733694, 0.820585921380499, 0.820585921380499, 
0.820585921380499, 0.820585921380499, 0.820585921380499, 0.820585921380499
), Cluster = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1)), row.names = c(NA, 
-10L), class = c("tbl_df", "tbl", "data.frame"))

question from:https://stackoverflow.com/questions/65624243/using-summarise-to-count-the-number-of-times-the-min-value-is-repeated

Answer 1

Your first option should ideally work but again floating point comparisons are not accurate. (Ref Why are these numbers not equal?)

Try rounding the numbers before using sum.

summarise(
  ...
  N_min = sum(round(Reachability, 2) == round(min(Reachability,na.rm = TRUE), 2))
  ...
)