Example: Data Frame Basics

Data frames are one of the most heavily used data structures in R.

Creation of a data frame

A data frame is created from scratch by supplying vectors to the data.frame function. Here are some examples:

x <- c(2.5, 3.5, 3.4)
y <- c(5, 10, 1)
my_df <- data.frame(x, y)
my_df

##     x  y
## 1 2.5  5
## 2 3.5 10
## 3 3.4  1

colnames(my_df) <- c("Floats", "Integers")

my_other_df <- data.frame(X = c(2, 3, 4), Y = c("A", "B", "C"))
my_other_df

##   X Y
## 1 2 A
## 2 3 B
## 3 4 C

The colnames function allows to supply column names to an existing data frame. Alternatively, the column names can be set within the data.frame function by assign the vector elements to a variable (capital X and Y in the example above).

Dimensions of a data frame

To get the dimensions of a data frame, use the ncol (number of columns), nrow (number of rows) or str (structure) function:

ncol(my_other_df)

## [1] 2

nrow(my_other_df)

## [1] 3

str(my_other_df)

## 'data.frame':	3 obs. of  2 variables:
##  $ X: num  2 3 4
##  $ Y: Factor w/ 3 levels "A","B","C": 1 2 3

Displaying and accessing the content of a data frame

The content of a data frame is accessed by either a position information given in square brackets (e.g. df[3,4]) or a column name given after a $ sign (e.g. df$columnName). Here is an example:

my_other_df[1,]  # Shows first row

##   X Y
## 1 2 A

my_other_df[,2]  # Shows second column

## [1] A B C
## Levels: A B C

my_other_df$Y  # Shows second column

## [1] A B C
## Levels: A B C

If position information is used, the ordering matters. If you think of a data frame like a table, then the following applies:

  • In a 1-D data frame, the first dimension is the row
  • In a 2-D data frame, the first dimension is the row, the second the column

Higher dimensions follow the same logic.

Here are some possible combinations:

  • Single row, all columns: df[x,]
  • Single column, all rows: df[,y]
  • Single row and column: df[x,y]
  • All except one row, all columns: df[-x,y]
  • Selected rows, all columns: df[c(x1, x2, x3),]
  • Continous rows, all columns: df[c(x1:x2),]

In summary, dimensions like rows or columns which should be selected have positive numbers, dimensions that should be hidden have negative numbers, and if all entries of a dimension should be selected one just leaves the field empty. If more than one dimension should be shown or hidden, one has to supply these information with a vector defined by the c function.

my_other_df[c(1,3),]  # Shows rows 1 and 3

##   X Y
## 1 2 A
## 3 4 C

my_other_df[c(1,2),]  # Shows rows 1 to 2

##   X Y
## 1 2 A
## 2 3 B

If you are interested in the first or last rows, you can also use the head or tail functions. The default number of lines to be displayed is five but this can be changed with the second argument. Let us have a look at the first two rows:

head(my_other_df, 2)

##   X Y
## 1 2 A
## 2 3 B

And now on the last two rows:

tail(my_other_df, 2)

##   X Y
## 2 3 B
## 3 4 C

Changing, adding or deleting an element of a data frame

In order to change an element of a data frame (individual value or entire vectors like rows or columns), you have to access it following the logic above. To add or delete a column, you have to supply/remove a vector to the specified position.

Other more specific changes will be covered later.

# overwrite an element
my_other_df$X[3] <- 400  # same as my_other_df[3,1] <- 400
my_other_df

##     X Y
## 1   2 A
## 2   3 B
## 3 400 C

# change an entire dimension
my_other_df[,1] <- c("200", "300", "401")  # same as my_other_df$X <- 400
my_other_df

##     X Y
## 1 200 A
## 2 300 B
## 3 401 C

# add a new column
my_other_df$z <- c(255, 300, 100)
my_other_df

##     X Y   z
## 1 200 A 255
## 2 300 B 300
## 3 401 C 100

# delete a column
my_other_df$z <- NULL
my_other_df

##     X Y
## 1 200 A
## 2 300 B
## 3 401 C

As for lists, to actually delete an element, it has to be set to NULL.

For more information have a look at e.g. the respective data type site at Quick R. There you will also find an overview on how to get information about an object. Of course, looking into the package documentation or search the web is always a good idea, too.

Updated: