Filter data frame rows based on values in vector

Question

What is the best way to filter rows from data frame when the values to be deleted are stored in a vector? In my case I have a column with dates and want to remove several dates.

I know how to delete rows corresponding to one day, using !=, e.g.:

m[m$date != "01/31/11", ]

To remove several dates, specified in a vector, I tried:

m[m$date != c("01/31/11", "01/30/11"), ]

However, this generates a warning message:

Warning message:
In `!=.default`(m$date, c("01/31/11", "01/30/11")) :
longer object length is not a multiple of shorter object length
Calls: [ ... [.data.frame -> Ops.dates -> NextMethod -> Ops.times -> NextMethod

What is the correct way to apply a filter based on multiple values?

score 40 · Accepted Answer · edited May 23 '17 at 12:06

40

nzcoops is spot on with his suggestion. I posed this question in the R Chat a while back and Paul Teetor suggested defining a new function:

`%notin%` <- function(x,y) !(x %in% y)

Which can then be used as follows:

foo <- letters[1:6]

> foo[foo %notin% c("a", "c", "e")]
[1] "b" "d" "f"

Needless to say, this little gem is now in my R profile and gets used quite often.

edited May 23 '17 at 12:06

Community

1
1

answered Sep 21 '11 at 11:17

Chase

67,710
18
144
161

Nice. I always find it breaks my concentration when I have to type something like `!(x %in% y)`... great tip. – Peter M Sep 21 '11 at 23:01
2

I don't think `dplyr` can handle this, e.g., `filter(df, foo %notin% c("a", "c", "e"))` – tumultous_rooster Sep 17 '15 at 22:13
1

It was just a comment! I was trying to use `%noin%` in a `filter` call and it didn't work, non-suprisingly. – tumultous_rooster Sep 18 '15 at 18:51
@MattO'Brien, I posted an equivalent using `anti_join` below . . . – Ben G Jul 05 '18 at 15:23

nzcoops · Answer 2 · 2011-09-21T23:58:44.417

14

I think for that you want:

m[!m$date %in% c("01/31/11","01/30/11"),]

edited Sep 21 '11 at 23:58

answered Sep 21 '11 at 05:15

nzcoops

9,132
8
41
52

This doesn't seem to work when date is in `as.POSIXct` class ! – vagabond Nov 21 '14 at 16:11

score 4 · Answer 3 · answered Apr 11 '13 at 10:30

4

cool way is to use Negate function to create new one:

`%ni%` <- Negate(`%in%`)

than you can use it to find not intersected elements

answered Apr 11 '13 at 10:30

Vova Naumov

51
1

score 2 · Answer 4 · answered Jul 05 '18 at 15:22

In regards to some of the questions above, here is a tidyverse compliant solution. I used anti_join from dplyr to achieve the same effect:

library(tidyverse)

numbers <- tibble(numbers = c(1:10))
numbers_to_remove <- tibble(number = c(3, 4, 5))

numbers %>%
  anti_join(numbers_to_remove)

Filter data frame rows based on values in vector

4 Answers4

Linked

Related