Removing selected observations from a dataframe in r

Question

I'm looking to remove 7 rows from a large dataset (>400 rows), based on the values in a certain column. I am having issues with this simple endeavour.

 ##Generate sample dataset
    Site.Num=c(1:20)
    Year=c(1990:2009)
    Day=c(10:29)
    Final<-data.frame(Site.Num,Year,Day)


 ##I would like to remove 5 rows, based on 5 sites from the Site.Num column
     Final <- Final[which(Final$Site.Num!=c(1,4,10,11,14)), ]


##I receive this error message
    Warning message:
        In Final$Site.Num != c(1, 4, 10, 11, 14) :
        longer object length is not a multiple of shorter object length

Answer 1

The warning is because you're using != to compare different vectors, and recycling will happen. However, this warning is important, because in this case, you're asking for a different value than you expect.

For example (using == for clarity) if you want to see which values of c(1,2,2) are contained in c(1,2) , consider this expression:

> c(1,2,2) == c(1,2)
[1]  TRUE  TRUE FALSE
Warning message:
In c(1, 2, 2) == c(1, 2) :
  longer object length is not a multiple of shorter object length

but 2 is clearly in both vectors. The FALSE value is because the vector on the right is being recycled, so these are the actual values compared:

> c(1,2,2) == c(1,2,1)
[1]  TRUE  TRUE FALSE

However, in the former case, the vector on the right is not recycled an integral number of times. This usually means that you did something that you didn't expect. You want the operator %in% which gives set inclusion:

> c(1,2,2) %in% c(1,2)
[1] TRUE TRUE TRUE

No warning, and the expected answer.

For your question, here is the command to get the desired rows:

Final <- Final[!(Final$Site.Num %in% c(1,4,10,11,14)), ]

Note that which doesn't help or hurt in this statement, unless the set of returned rows would be empty.

Answer 2

使用 dplyr 包，你可以做这样的事情。

filter(Final, !Site.Num %in% c(1,4,10,11,14))