Basic Data Structures III: Data Frames in R

What are Data Frames in R?

Data Frames in R are a two dimensional data structure used to hold data of heterogeneous types in a table-like format.

A data frame comprises of rows and columns. Columns are the variables that you record in an experiment. Rows are the observation. More specifically, each row is an instance of an observation.

Another way of looking at data frame is to view it as a matrix that can hold data of various different data types. That is to say that each row can have data of different types, but each column has a data of the same type.

Yet another way of perceiving data frame is to view them as a list of vectors or to put simply as a collection of vectors that are cbound (i.e. cbind operation has been applied on them).

The following code snippet shows the structure of a data frame called mtcars that ships with R-

Creating a Data Frame in R

Like with other data structures in R – there are two interconnected ways of creating data frames –

1. Using the ‘constructor’ data.frame function
2. as.data.frame function (this is internally called by data.frame function)

Creating Data Frame using data.frame Function

Let’s create a data frame that we will be using for illustration in the rest of article. We will record data about our favorite lowlifes viz. politicians. We want to record their names (which I will hide for my security), the last criminal offence they committed and when was their last conviction in a court of law.  I agree this is going to be a pretty nasty example, but there is a message – name and shame politicians when you can.

Let’s first prepare the data, we will then pump all of it into a data frame.

Note that all three vectors have the same length because these three vectors hold data pertaining to five different low-lifes err… politicians. Our next task is to hook these three scattered vectors into a single data frame. We do that by using data.frame function as shown below –

The only thing new in the above code is stringsAsFactors argument to data.frame. This logical value should be set to FALSE otherwise R will automatically convert character data to factors.

Creating Data Frames with Non-Default Row Names

Among other arguments of data.frame function, there is one that you will probably use quite often and that is row.names . Using this argument you can pass in a values or a vector that holds a distinct, unique name for each of the row of your data.

Instead of integers as row numbers previously, we are now using the names of politicians as the row name. (Sympathies for the row)

Now we have a data frame called lowlifes that holds all the data about these …, well, you get the point.

Of course, data.frame can be used to created data frames of much more socially useful and better things. Otherwise there would be no sense in learning about them (and I mean data frames).

Creating Data Frame using as.data.frame Generic Function

as.data.frame function can be used to transform another data structure such as vector, matrix to a data frame as demonstrated below –

Characteristic Features of Data Frames in R

The following are defining features of data frames –

1. Every row of data frame has a unique name. Duplicate row names are not allowed.
2. No row can have its name missing i.e. NA is not a valid row name.
3. Row names must be of integer or character data types. You cannot, for example, have logicals as row names.
4. If you don’t explicitly provide row names, R uses row numbers starting with integer 1 as row names.
5. NULL is a valid for a row name, Using NULL for the value resets the row names to seq_len(nrow(data frame)). Or in other words the rows will be named 1..len(rows).
6. The column names should be non-empty. It’s technically allowed to have duplicate column names, but quite often doing that makes no sense at all. For example, during an observation you don’t want to have two variables to record whether a politician is also a criminal (anyway, the outcome is a foregone conclusion).
7. You can skip providing column names explicitly, in which case your column names will be a concatenation of column values.

Accessing Members of a Data Frame

You can access members of a data frame in a number of ways shown in next few sections –

Accessing Single Element By Its Address

You can use the position of a member to access it in a data frame –

Similarly, you can use name of elements.  For example, you can find out about the atrocities of the 3rd politician as follows as well –

Accessing Rows From a Data Frame

Using the ‘[]‘ operator, you can select entire rows.

You can, of course, use names of rows and columns instead of their positions –

Accessing Columns From a Data Frame

Different ways of extracting columns mirror the ways of extracting rows.

Continuing our journey into knowing more about the illustrious politicians around us, we will now extract their ‘attributes’ – (Pun intended)

Using ‘[‘

Using names

Retrieving Columns of Data Frames in R

Suppose that you are so enamored by the good deeds of politicians that you want to gather all their wonderful deeds as a vector for further anlysis, exploration and citation. R can help you in this charitable act in three different ways –

1. Extracting Column Vector Using $ Operator

There you go, R gives a character vector with all the details.

2. Extracting Column Vector Using [ Operator

3. Extracting Column Vector Using [[ Operator

‘[[‘ is a source of perpetual pain in R. But you can use it with column number or name to obtain a vector –

Operations on Data Frames

There are a number of operations permitted on data frames –

Transpose of a Data Frame

As indicated in the opening paragraph, a data frame can be viewed as a matrix of heterogeneous data type elements. You can calculate transpose of a data frame using the t function. Though transpose of a data frame does not make much mathematical sense, it can be used to swap rows with columns.

Adding Columns to Data Frame

Suppose that you also want to record the number of pending criminal cases of each, well, politician. No hassles. You can use the $ operator even after the data frame is constructed.

Removing Columns from a Data Frame

Removing a column from a data frame is equally easy. Using the $ operator you set the column to NULL –

Adding Rows to Data Frame

Suppose you have the inner urge to model more politicians and add them to your data frame. R makes that pretty straightforward. In the example here we create a new data frame called more.rascals and rbind it to the existing data frame df2

Deleting Rows from a Data Frame

Suppose some politician P6 threatens you and you agree to remove his details from lowlives data frame. Deleting rows can be accomplished by using the – operator as shown below –

Sorting a Data Frame

You can sort a data frame based on values in one or more columns using the order() function. Let’s sort the lowlives data frame based on the alphabetical ordering of the offences that they have committed.

Changing Order of Rows or Columns in a Data Frame

The order of columns and rows in a data frame can be altered using various selection mechanisms discussed in the beginning of this article. Consider the trivial example below –

Conclusion

This article introduced one of the most frequently used data structures in R, viz. data frames. We have already discussed vectors and matrices. We will look at lists in R next.

 

We are social

Spread the word
Follow CodingRaptor