R Sort DataFrame Rows by Column Value

How to sort data frame in R? To sort data frame by column values use the order() function. By default, the sorting function performs in ASCENDING order and provides an option to sort in descending order. Also, by default, all NA values on the sorting column are kept at the last and you can change this behavior by using optional params.

1. Quick Examples of Sorting DataFrame

Following are quick examples of how to sort DataFrame by column value in ascending and descending order in R programming.


# Below are the quick examples of sorting data frame

# Example 1: Sort data frame by specific column
df2 <- emp_df[order(df$price),]

# Example 2: Sort by multiple columns
df2 <- df[order(df$price,df$name ),]

# Example 3: Sort descending order
df2 <- df[order(df$price,decreasing=TRUE),]

# Example 4: Sort by putting NA top
df2 <- df[order(df$price,decreasing=TRUE, na.last=FALSE),]

# Example 5: Load dplyr library
library(dplyr)
df2 <- df %>% arrange(price)
df2 <- df %>% arrange(desc(price), desc(name) ))

# Example 6: Load data.table library
library("data.table")
df2 <- setorder(df,price)

Let’s create an R DataFrame with the columns of "id", "name", "price", and "publish_date". Apply different sorting methods to the data frame and get the new data frame by reordering the rows based on values of single/multiple columns.


# Create Data Frame
df = data.frame(id=c(11,22,33,44,55),
          name=c("spark","python","R","jsp","java"),
          price=c(144,NA,321,567,567),
          publish_date= as.Date(
            c("2007-06-22", "2004-02-13", "2006-05-18",
              "2010-09-02","2007-07-20"))
          )
df

Yields below output.

2. Sort DataFrame in R using order() Function

The order() is a base function that is used to sort the data frame in R based on column value, this function can also be used to sort vectors. It takes the ordered column indices, so we have to use [] – index, and inside this, we can apply the order() function. Hence this will return the column.

2.1 Syntax of order() Function

Following is the syntax of the order() function.


# order() syntax
order(data, na.last = TRUE, decreasing = FALSE)

data – vector or command separated vectors to sort on multiple columns.
na.last – Default set to TRUE. Use FALSE to have all NA values put at the top. Use NA to remove.
decreasing – Default set to FALSE. Use TRUE to order in descending.

2.2 Example of Sort DataFrame by Column Value

Let’s apply the order() function to sort the R dataframe by column value in ascending order. The following example sorts the data by column price. Since this function takes the vector as an argument, use df$price it as an argument. Note that in R, every column in DataFrame is a vector.


# Sort DataFrame by specific column value
df2 <- df[order(df$price),]
df2

Yields below output.

4. By Multiple Columns

If you want to sort by multiple columns, you can pass all columns/variables into the order() function. It will return the new data frame by re-ordering the rows of the original data frame df based on specified columns.

In this case, Pass price and name columns into the order() function. It sorts first by price and then by name.

The result is a new data frame df2 with rows arranged in ascending order based on the price column. In case of ties in the price, the rows are further sorted based on the name column.


# Sort by multiple columns
df2 <- df[order(df$price, df$name ),]
df2

# Output:
#   id   name price publish_date
# 1 11  spark   144   2007-06-22
# 3 33      R   321   2006-05-18
# 5 55   java   567   2007-07-20
# 4 44    jsp   567   2010-09-02
# 2 22 python    NA   2004-02-13

6. Sort DataFrame by Descending order

By default, sorting happens in ascending or increasing order, for string columns, the sort happens in alphabetical order (A to Z). You can sort the data.frame in descending order by using decreasing=TRUE.


# Sort descending order
df2 <- df[order(df$price,decreasing=TRUE),]
df2

# Output:
#   id   name price publish_date
# 1 11  spark   144   2007-06-22
# 3 33      R   321   2006-05-18
# 2 22 python    NA   2004-02-13
# 4 44    jsp   567   2010-09-02
# 5 55   java   567   2007-07-20

When arranging the rows you can also prefix with a minus sign to perform sorting in decreasing order. This allows you to sort one column in ascending and another column in descending order.

7. Sort by having NA First

By default rows with NA on a sorting column will be put at the end in R, however, you can change this behavior and put rows with NA on top of the dataframe by using na.last=FALSE.


# Sort by putting NA at first
df2 <- df[order(df$price,decreasing=TRUE, na.last=FALSE),]
df2

# Output:
#   id   name price publish_date
# 2 22 python    NA   2004-02-13
# 4 44    jsp   567   2010-09-02
# 5 55   java   567   2007-07-20
# 3 33      R   321   2006-05-18
# 1 11  spark   144   2007-06-22

8. Sort DataFrame using dplyr Package

arrange() function from dplyr package is also used to arrange the values in an ascending or descending order. To use the arrange() function, you have to install dplyr first using install.packages(‘dplyr’) and load it using library(dplyr).

All functions in dplyr package take data.frame as a first argument. When we use dplyr package, we mostly use the infix operator %>% from magrittr, it passes the left-hand side of the operator to the first argument of the right-hand side of the operator. For example, x %>% f(y) converted into f(x, y) so the result from the left-hand side is then “piped” into the right-hand side.


# Load dplyr library
library(dplyr)
df2 <- df %>% arrange(price)
df2

To sort columns in descending order


# Sort descending order
df2 <- df %>% arrange(desc(price))
df2

9. Sort DataFrame using data.table Package

You can also use the setorder() function from the data.table to perform sorting on DataFrame columns. This function takes the data.frame object and column as input and return a new DataFrame after sorting by the specified column.


# Load data.table library
library("data.table")
df2 <- setorder(df,price)
df2

Conclusion

In this article, I have explained how to use the order() function to sort the DataFrame values by single/multiple columns in R and how to sort by descending order, keeping all NA values first. Finally converted using the arrange() function from the dplyr package and the setorder() function from the data.table to order data.frame by specified columns.

Similarly, you can also sort by date column, when performing an order on a date make sure the column is in date type.

References

https://www.rdocumentation.org/packages/base/versions/3.6.2/topics/order