如何根据R中另一个相同的行值对行百分比求和？

Question

My dataset is composed of whale calls. I have two variables: nclicks and percent. nclicks refers to the number of clicks in the call (ranges from 3 to 30). Percent refers to the frequency with which that type of call was made in a given year. I would like to sum the percentages for longer calls (those with 11+ clicks) and add a new row to the dataframe that has 11+ for nclicks and the summed percentage for percent. I then want to delete the rows that made up the new row.

I've tried coding "nclicks" as both a factor and a numeric. I've used combinations of aggregate, rowSums, rbind, etc. but with no luck. The closest I've come was getting a new row that had the summed percentages, but I had to specify which rows to include manually (see example below). This method also summed the nclicks values (so in my example below, I get a new row with 43 (11+12+20) in nclicks and 20 in percent, when I really want the row number to be 4, nclicks to be 11+, and percent to be 20).

nclicks=c(3,4,5,11,12,20) 

percent=c(30,30,20,10,5,5) 

df=data.frame(cbind(nclicks,percent)) 

df["11+",]=df["4",]+df["5",]+df["6",] 

df=df[-c(4,5,6), ] 

df

This is what I end up with:

 nclicks percent
1         3      30
2         4      30
3         5      20
11+      43      20

I want to sum the percentages of rows for which the value of nclicks is > 10, but I'm having trouble executing this. I don't want to have to individually specify which values of nclicks to include, because some years have many different nclick values > 10 while some years only have a few different values > 10.

Answer 1

You can create a group column to help aggregate the rows where nclicks>=11 .

library("tidyverse")

nclicks <- c(3, 4, 5, 11, 12, 20)
percent <- c(30, 30, 20, 10, 5, 5)

df <- tibble(nclicks, percent)
df <- df %>%
  mutate(group = ifelse(nclicks >= 11, "11+", nclicks)) %>%
  group_by(group) %>%
  summarise_at(vars(nclicks, percent), sum)
df
#> # A tibble: 4 x 3
#>   group nclicks percent
#>   <chr>   <dbl>   <dbl>
#> 1 11+        43      20
#> 2 3           3      30
#> 3 4           4      30
#> 4 5           5      20

^{Created on 2019-03-31 by the reprex package (v0.2.1)}