Data are read into R using several functions written for this purpose. In my experience .csv format is the most well behaved.
The read.csv() function is used to assign a data file to
an object:
data <- read.csv("the path and
file.csv",other arguments here)
Let’s take a look at the help: ?read.csv.
There is a more general function, read.table, that can
read different types of files. But, it requires more programming. In
addition, there is a package called foreign that has
functions for reading files from other programs
(e.g. read.spss reads SPSS files; read.dta
reads Stata files, etc.). However, in my experience, if you are sending
something from Stata to R, for example, it is better to just have Stata
write a .csv file using: the outsheet function in Stata
(type help outsheet in Stata for information).
Let’s work through an example. Start by downloading the “r-workshop-data.csv” file from this link on the RWorkshop website.
The file contains 52 individuals and 4 variables. The variables are: respondents id (“id”), a binary variable indicating whether the respondent is male or female (“male” where “1” is male), a measure of the respondent’s age (“age”), and a measure of risky behaviors engaged in by the respondent (“risky”).
Let’s go ahead and import it into R:
First, set the directory where the file is by using the
setwd("/…") function. If you’re a windows user, replace
every "\" in the file path with "\\".
#read in the data file.
dat <- read.csv(
"r-workshop-data.csv", # the data file.
header = TRUE, # tell R to read the first row as variable names.
as.is = TRUE, # tell R to not make any conversions.
na.strings = "." # tell R that missing values are periods.
)
dat #look at the object.If you had a REALLY big data set, you could use the
head() function or the tail() function to look
at the first few (or last few) cases.
If you use the View() function, you can open the
spreadsheet to see how R views the file and why we give it particular
instructions.
Note that we could skip the step of storing the .csv file locally and simply call file from the website url. For example:
dat <- read.csv(
"https://raw.githubusercontent.com/jacobtnyoung/RWorkshop/main/r-workshop-data.csv",
header=TRUE, as.is=TRUE, na.strings="." #all the other arguments remain the same.
)
dat #look at the object.Now, we can work with the dat object. As we saw before
with matrices, we can index parts of the object by referring to the
dimensions:
dim( dat ) # what are the dimensions of the object?
dat[,1] # first column or first variable.
dat[1,] # first row or first case.
dat[,c( 1,2 )] # first two columns.
dat[,1:2] # same thing as dat[,c(1,2)].
dat[1:25,] # just the first 25 cases.
dat[,-c(1)] # remove the first column.We can also refer to the variables (i.e. columns) by their names
enclosed in "" or by using the $ key.
dat[,"id"] # returns the entire column named "id".
dat$id # returns the column called "id" within the object "dat".Referencing parts of an object in this way makes it easy to execute
functions. For example, the function summary() provides the
range, mean, median, and info about missing values:
summary( dat ) #summarize all the variables.
summary( dat$age ) #just summarize age.
summary( dat[,"age"] ) #same thing, different syntax.
summary( dat[,c( "age","male" )] ) #just age and male.Let’s read into the data, but exclude the header to see how to assign names:
#toggle the header argument by setting it to FALSE.
dat <- read.csv( "r-workshop-data.csv" , header = FALSE, as.is = TRUE)
dat # look at the object, note the first row.
dat[1,] # R has created the variable names for us.
#we can see this from the colnames() function.
colnames( dat )
#change the names.
names <- c( "id","male","age","risky" ) # create an object of names.
colnames( dat ) <- names #tell R what the column names are.Obviously, if we had the names we would read them in with
header=TRUE in the read.csv line.
We can merge two objects (or datasets) with the merge()
function. For example, let’s create two datasets where the cases do not
match exactly (i.e. some cases are missing data on a merged variable) by
using the data.frame() function:
# Create the datasets (note the difference between them).
dat.a <- data.frame( id = c( 1,2,3,4,5 ), age = c( 10,12,14,15,11 ) )
dat.b <- data.frame( id = c( 1,2,3,5 ) , sex = c( "m","m","f","m" ) )
class( dat.a ) # object "data.a" is of class data.frame.
?merge # note the different merging options.
# We can now merge these two datasets to get a single object.
dat.c <- merge( dat.a,dat.b,by.x = "id" )
dat.d <- merge( dat.a,dat.b,by.x = "id", all.x=TRUE ) #note difference.We can also write data files out for analysis in another program. For
example, say we have created a variable in our data set that we could
only create in R and we want to export the file to another program. For
.csv files, we can just use the write.csv function. As with
reading files, there is a more general function,
write.table, that can write different types of files (see
?write.table). For write.csv we just tell it
what object we want to write out to a file and the file name:
First, set the directory where the file will go by using
setwd("/…"). Then:
Recall that to save the workspace use the save.image()
function. This function requires a file path, a file name, and the
extension “.RData” which is the format for an R workspace file. Since
our imported data is an object (i.e. data), it will be
saved to the workspace:
First, set the directory where the file will go by using
setwd("/…"). Then:
save.image( "dat.RData" ) #file path can be added also.
rm( list=ls() ) # clear the workspace.
ls() # check the contents of the workspace.
load( "dat.RData" ) #load the workspace that was previously saved.