R：根据文件名列从数据框中删除行

Question

I've got a large csv file I'm reading into a data frame which is itself a combination of csv's. The first column in the data frame is the file name. The file name always ends with a 5 digit number and ".csv" The number of occurrences of each file name will vary. Ex:

Source File
xxx_00001.csv
xxx_00001.csv
xxx_00001.csv
xxx_00001.csv
xxx_00001.csv
xxx_00002.csv
xxx_00002.csv
xxx_00002.csv
xxx_00002.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
...

How would I go about removing the rows associated with the last n occurrences of the file name? (Say, the last 2?) I'd like to end up with:

Source File
xxx_00001.csv
xxx_00001.csv
xxx_00001.csv
xxx_00002.csv
xxx_00002.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
xxx_00003.csv
...

Answer 1

Using dplyr :

 library(dplyr)
 n_to_remove <- 2
 filtered <- group_by(df, SourceFile) %>% slice(1:(n()-n_to_remove))

group_by will make sure that the slice operation occurs for each group separately. n() is also a function from dplyr that will return the number of rows inside the group. Note that this will fail if the number of rows of one of the CSVs are smaller than n_to_remove .

Answer 2

We can use ave from base R

n <- 2
df1[with(df1, !ave(seq_along(Source_File), Source_File, 
             FUN = function(x) x %in% tail(x,n))), , drop=FALSE]
#     Source_File
#1  xxx_00001.csv
#2  xxx_00001.csv
#3  xxx_00001.csv
#6  xxx_00002.csv
#7  xxx_00002.csv
#10 xxx_00003.csv
#11 xxx_00003.csv
#12 xxx_00003.csv
#13 xxx_00003.csv

---

Or with data.table

library(data.table)
setDT(df1, keep.rownames=TRUE)[, head(.SD, -n) ,.(Source_File)][, rn:=NULL][]
#     Source_File
#1: xxx_00001.csv
#2: xxx_00001.csv
#3: xxx_00001.csv
#4: xxx_00002.csv
#5: xxx_00002.csv
#6: xxx_00003.csv
#7: xxx_00003.csv
#8: xxx_00003.csv
#9: xxx_00003.csv

data

df1 <- structure(list(Source_File = c("xxx_00001.csv", "xxx_00001.csv", 
"xxx_00001.csv", "xxx_00001.csv", "xxx_00001.csv", "xxx_00002.csv", 
"xxx_00002.csv", "xxx_00002.csv", "xxx_00002.csv", "xxx_00003.csv", 
"xxx_00003.csv", "xxx_00003.csv", "xxx_00003.csv", "xxx_00003.csv", 
"xxx_00003.csv")), .Names = "Source_File", class = "data.frame", 
row.names = c(NA, -15L))