Removing rows from a df based on the values of multiple columns

Question

I want to remove from this df rows where m=f, and where t is within 4 of rows where that happen whatever the specific combination of s and b is.

s<- c(1,1,1,1,1,1,2,1)
b<- c(1,1,1,1,1,1,1,2)
m <- c("o","o","o","o","f","o","o","o")
data <- rnorm(8)
t <- c(1,2,3,4,5,7,8,9)
df<- as.data.frame(cbind(s,b,m,data,t))

So in this example I want to remove the row containing 'f' and the rows which contain t of value up to 4 more than this row (so t of 5 to 9), provided that their value of s and b is the same as the row where 'f' was. So in it should remove row 5 and 6, but not row 7 because s is different from where 'f' was detected and not row 8 because b is different.

Sorry it is quite a dense question, I haven't been able to formulate an attempt at it.

vaettchen · Accepted Answer · 2014-01-30 08:21:08Z

It is generally regarded good practice not to to use function names for variables, so I re-christen your sample data to myDF. I also assume that you don't really want all your columns to be factors what is what you get using cbind() (check out with str( df ). Lastly, it is probably a good idea to make sure that the random data remain consistent. Hence:

set.seed( 55 )
myDF <- data.frame( s = c( 1, 1, 1, 1, 1, 1, 2, 1 ),
                    b = c( 1, 1, 1, 1, 1, 1, 1, 2 ),
                    m = c( "o", "o", "o", "o", "f", "o", "o", "o" ),
                    data = rnorm( 8 ),
                    t = c( 1, 2, 3, 4, 5, 7, 8, 9 ) )

Here is how I would address your problem - I'm sure there are much better ways but this should at least work:

n <- which( myDF$m == "f" )
m <- which( myDF$t >= myDF$t[n] & myDF$t <= myDF$t[n] + 4 )
p <- which( myDF$s == myDF$s[n] )[m]
o <- which( myDF$b == myDF$b[n] )[p]
o <- o[ !is.na( o ) ]
myDF <- myDF[ -o, ]

Collectives™ on Stack Overflow

Removing rows from a df based on the values of multiple columns

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related