What's your favorite one-liner in R?
Include a short companion example, and limit to one tip per post, please.
Note, ; is cheating.
Example: calculate x[i] / x[i-1] for a vector x,
x <- 1:10
Reduce("/", as.data.frame(embed(x, 2)))
(collected from R-help, I forget who/when)
Edit: after some initial controversy, it looks like the question is now reopen for entries.
If you want to record the time that you created a file in its name (perhaps to make it unique, or prevent overwriting), then try this one-line function.
timestamp <- function(format = "%y%m%d%H%M%S")
{
strftime(Sys.time(), format)
}
Usage is, e.g.,
write.csv(
some_data_frame,
paste("some data ", timestamp(), ".csv", sep = "")
)
Get odd or even indices.
odds <- function(x) seq_along(x) %% 2 > 0
evens <- function(x) seq_along(x) %% 2 == 0
Usage is, e.g.,
odds(1:5)
evens(1:5)
I often need fake data to illustrate, say, a regression problem. Instead of
X <- replicate(2, rnorm(100))
y <- X[,1] + X[,2] + rnorm(100)
df <- data.frame(y=y, X=X)
we can use
df <- transform(X <- as.data.frame(replicate(2, rnorm(100))),
y = V1+V2+rnorm(100))
to generate two uncorrelated predictors associated to the outcome y.
Removing NaNs - which are a nuisance every once in a while - from a vector or dataframe (found somewhen on R-help)
is.na(x) <- is.na(x)
Example:
> x <- c(1, NaN, 2, NaN, 3, NA)
> is.na(x) <- is.na(x)
> x
[1] 1 NA 2 NA 3 NA
Convert Excel dates to R dates. Answer adapted from code by Paul Murrell.
excel_date_to_r_date <- function(excel_date, format)
{
#excel_date is the number of days since the 0th January 1900. See
#http://www.stat.auckland.ac.nz/~paul/ItDT/HTML/node67.html
strftime(as.Date(as.numeric(excel_date) - 2, origin = "1900-01-01"), format)
}
Usage is, e.g.,
excel_date_to_r_date(40700, "%d-%m-%Y")
Not quite what you are after, but fitting a multivariate linear regression model in one line is great:
lm(y ~ x1 + x2)
Reduce() is a new kid on the block. The same can be done using do.call(), and is a little bit quicker (on my system at least):
do.call("/", as.data.frame(embed(1:10, 2)))
R> do.call("/", as.data.frame(embed(1:10, 2)))
[1] 2.000000 1.500000 1.333333 1.250000 1.200000 1.166667 1.142857 1.125000
[9] 1.111111
R> Reduce("/", as.data.frame(embed(1:10, 2)))
[1] 2.000000 1.500000 1.333333 1.250000 1.200000 1.166667 1.142857 1.125000
[9] 1.111111
Function summarize the amount of missing data for each variable in a data frame. Returns a list.
propmiss <- function(dataframe) lapply(dataframe,function(x) data.frame(nmiss=sum(is.na(x)), n=length(x), propmiss=sum(is.na(x))/length(x)))
Not a one-liner, but returning this info as a data frame is more useful.
propmiss <- function(dataframe) {
m <- sapply(dataframe, function(x) {
data.frame(
nmiss=sum(is.na(x)),
n=length(x),
propmiss=sum(is.na(x))/length(x)
)
})
d <- data.frame(t(m))
d <- sapply(d, unlist)
d <- as.data.frame(d)
d$variable <- row.names(d)
row.names(d) <- NULL
d <- cbind(d[ncol(d)],d[-ncol(d)])
return(d[order(d$propmiss), ])
}
Wipes the slate clean removes all objects from the memory.
rm(list=ls(all=TRUE))
Multiple columns edit is one of my favourite.
E.g. to change all numeric columns to characters:
X <- iris
X[id] <- lapply(X[id <- sapply(X, is.numeric)], as.character)
or standardize them
X[id] <- lapply(X[id <- sapply(X, is.numeric)], scales)
Well, not really a oneliner but textConnection is great!
x <- "1,3
1,a
1,g,4
3,d,6
2,X,1,3
2,K"
read.table(textConnection(x), sep=",", header=FALSE, na.strings="", fill=TRUE)
result
V1 V2 V3 V4
1 1 3 NA NA
2 1 a NA NA
3 1 g 4 NA
4 3 d 6 NA
5 2 X 1 3
6 2 K NA NA
>
Here's another tip collected from R-help (if memory serves, by Romain François).
Remove existing variables from the workspace:
rm( list = Filter( exists, c("a", "b") ) )
My favorite one-liner can be found in the help pages of the %in% function and is basically its opposite.
f.wo <- function(x, y) x[!x %in% y]
Wrapped up into a nice, small function it comes really handy. E.g.
R> f.wo(c("a", "b", "c"), "b")
[1] "a" "c"
R> f.wo(1:8, c(2,7))
[1] 1 3 4 5 6 8
Function to read space delimited data from the clipboard
read.cb <- function(...) read.table(file="clipboard", ...)
e.g.
# read data from the clipboard with a header
d<-read.cb(T)
#read data from clipboard without header
d<-read.cb()
Function to convert columns of data in a data frame to factor variables
factorcols <- function(d, ...) lapply(d, function(x) factor(x, ...))
E.g. convert columns 1-4 in data frame d to factor variables
d[1:4] <- factorcols(d[1:4])
Return a new matrix, where the rows of the original matrix are sorted by columns:
newmat <- t(apply(orimat, 1, sort))
