是否有任何 function 从 R 的数据框中提取几行（不连续）？

Question

I am trying to extract several rows (not one after the other) from data frame. The range of rows that should be removed is in another data frame.

I have tried to extract the rows with for loop, but unfortunately only the last range is removed.

this is the code line I used (inside of 'for' loop, while 'i' is the variable):

new_df <- main_df[-(erase_df$starts[i]:erase_df$stops[i]),]

for example: this is the data frame that I want to change (main_df)

> main_df
   v1  v2     v3
1   1 bla blabla
2   2 bla blabla
3   3 bla blabla
4   4 bla blabla
5   5 bla blabla
6   6 bla blabla
7   7 bla blabla
8   8 bla blabla
9   9 bla blabla
10 10 bla blabla
11 11 bla blabla
12 12 bla blabla
13 13 bla blabla
14 14 bla blabla
15 15 bla blabla

this is the data frame (erase_df) that include the ranges of rows that I want to remove ('starts' vector indicate the first range of rows should be removed and 'stops' vector indicates the last row should be removed in that range)

> erase_df
  starts stops
1      3     5
2      9    10
3     12    14

so the new data frame should look like this:

> new_df
   v1  v2     v3
1   1 bla blabla
2   2 bla blabla
6   6 bla blabla
7   7 bla blabla
8   8 bla blabla
11 11 bla blabla
15 15 bla blabla

I expected output looks like (new_df) as I mentioned above, but instead of it, only the last range from erase_df has removed (starts = 12, stops = 14)

Answer 1

If you Map the seq function over erase_df to create a sequence of rows to remove for each row, then unlist them all into a single vector, you can subset main_df with the negative of that vector to remove the rows in the given ranges.

remove <- unlist(Map(seq, erase_df[[1]], erase_df[[2]]))

main_df[-remove,]
#    v1  v2     v3
# 1:  1 bla blabla
# 2:  2 bla blabla
# 3:  6 bla blabla
# 4:  7 bla blabla
# 5:  8 bla blabla
# 6: 11 bla blabla
# 7: 15 bla blabla

Or, for a more complex option that could be more efficient with larger data (haven't tested, just guesssing)

library(data.table)
setDT(main_df)
setDT(erase_df)

setkey(erase_df, starts, stops)
main_df[, v0 := v1]
for_anti <- 
  foverlaps(main_df, erase_df, by.x = c('v0', 'v1'), type = 'within',
            nomatch = NULL)

main_df[!for_anti, on = .(v1)]
#    v1  v2     v3 v0
# 1:  1 bla blabla  1
# 2:  2 bla blabla  2
# 3:  6 bla blabla  6
# 4:  7 bla blabla  7
# 5:  8 bla blabla  8
# 6: 11 bla blabla 11
# 7: 15 bla blabla 15