R - Replace NA values with 0 (zero) - Spark By {Examples} (2024)

How do I replace NA values on a numeric column with 0 (zero) in an R DataFrame (data.frame)? You can replace NA values with zero(0) on numeric columns of R data frame by using is.na(), replace(), imputeTS::replace(), dplyr::coalesce(), dplyr::mutate_at(), dplyr::mutate_if(), and tidyr::replace_na() functions.

For numeric columns, it is best to replace them with zero or any value that makes sense, and for strings, replace them with empty space. Using these methods you can also replace NA values with empty string.

Generally, NA values are considered missing values, and doing any operation on these values results in inconsistent results, hence before processing data, it is good practice to handle these missing values. In this article, we will see how to replace NA values with Zero in an R data frame with examples like replaced by a single index, multiple indexes, single column name, multiple column names, and on all columns.

1. Quick Examples of Replace NA Values with 0

Below are quick examples of how to replace data frame column values from NA to 0 in R.

#Quick Examples#Example 1 - Replace na values with 0 using is.na()my_dataframe[is.na(my_dataframe)] <- 0#Example 2 - Replace on selected columnmy_dataframe["pages"][is.na(my_dataframe["pages"])] <- 0print(df)#Example 3 - By using replace() & is.na()my_dataframe <- replace(my_dataframe, is.na(my_dataframe), 0)#Example 4 - Another waymy_dataframe <- my_dataframe %>% replace(is.na(.), 0)#Example 5 - Load the imputeTS packagelibrary("imputeTS")#Replace NA avalues with 0my_dataframe <- na_replace(my_dataframe, 0)#Example 6 - Replace NA with zero on all numeric columnlibrary("dplyr")my_dataframe <- mutate_all(my_dataframe, ~coalesce(.,0))#All below examples required these librarieslibrary("tidyr")library("dplyr")#Example 7 - Replace NA with zero on all numeric columnmy_dataframe <- mutate_all(my_dataframe, ~replace_na(.,0))#Example 8 - Replace NA using setnafill() from data.tablelibrary("data.table")my_dataframe <- setnafill(my_dataframe, fill=0)#Example 9 - Replace na with zero on specific numeric column#Load dplyr librarymy_dataframe <- my_dataframe %>% mutate(id = coalesce(id, 0))# Example 10 - Replace on multiple columnsmy_dataframe <- my_dataframe %>% mutate(id = coalesce(id, 0), pages = coalesce(pages, 0))# Example 11 - Load tidyr librarymy_dataframe <- my_dataframe %>% mutate_at(1, ~replace_na(.,0))# Example 12 - Replace NA on multiple columns by Indexmy_dataframe <- my_dataframe %>% mutate_at(c(1,3), ~replace_na(.,0))# Example 13 - Replace NA on multiple columns by namemy_dataframe <- my_dataframe %>% mutate_at(c('id','pages'), ~replace_na(.,0))# Example 14 - Replace only numeric columnsmy_dataframe <- my_dataframe %>% mutate_if(is.numeric, ~replace_na(., 0))

As you noticed above, I have used the following methods to replace NA values with 0 in R.

• Using is.na()
• Using replace()
• Using replace() from imputeTS package
• Using coalesce() from dplyr package
• Using mutate(), mutate_at(), mutate_if() from dplyr package
• Using replace_na() from tidyr package
• Using setnafill() from data.table package

Let’s create a data frame with some NA values, run these examples and validate the result.

#Create dataframe with 5 rows and 3 columnsmy_dataframe=data.frame(id=c(2,1,3,4,NA), name=c('sravan',NA,'chrisa','shivgami',NA), gender=c(NA,'m',NA,'f',NA))#Display dataframeprint(my_dataframe)

Output:

#Output id name gender1 2 sravan <NA>2 1 <NA> m3 3 chrisa <NA>4 4 shivgami f5 NA <NA> <NA>

2. Replace NA values with 0 using is.na()

is.na() is used to check whether the given data frame column value is equal to NA or not in R. If it is NA, it will return TRUE, otherwise FALSE. So by specifying it inside-[] (index), it will return NA and assigns it to 0. In this way, we can replace NA values with Zero(0) in an R DataFrame.

#Replace na values with 0 using is.na()my_dataframe[is.na(my_dataframe)] = 0#Display the dataframeprint(my_dataframe)

Output:

#Output id name gender1 2 sravan 02 1 0 m3 3 chrisa 04 4 shivgami f5 0 0 0

In the above output, we can see that NA values are replaced with 0’s.

3. Replace NA values with 0 in a DataFrame using replace()

Let’s see another way to change NA values with zero using the replace(). It will take three parameters.

#Replace NA avalues with 0my_dataframe <- replace(my_dataframe,is.na(my_dataframe),0)

1. the first parameter is the input data frame.
2. the second parameter takes is.na() method to check if it is NA
3. the last parameter takes value 0, which will replace the value present in the second parameter

Output:

# Output id name gender1 2 sravan 02 1 0 m3 3 chrisa 04 4 shivgami f5 0 0 0

In the above output, we can see that NA values are replaced with 0’s.

4. Replace NA values with 0 using replace() from “imputeTS”

replace() is used to replace NA with 0 in an R data frame. It is available in imputeTS package. so we have to install and load this package before using rename() method.

imputeTSis a third-party library hence, in order to use imputeTS library, you need to first install it by usinginstall.packages('imputeTS'). Once installation completes, load theimputeTSlibrary in order to use thisreplace()method. To load a library in R, uselibrary("imputeTS").

#Replace NA avalues with 0my_dataframe <- na_replace(my_dataframe, 0)

Output:

# Output id name gender1 2 sravan 02 1 0 m3 3 chrisa 04 4 shivgami f5 0 0 0

In the above output, we can see that NA values are replaced with 0’s.

5. Replace NA with Zero on All Numeric Values

There are several other ways to rename NA with zero in the R data frame by using methods from the dplyr package.

All previous examples use the Base R built-in functions that can be used on a smaller dataset but, for bigger data sets, you have to use methods from dplyr package as they perform 30% faster. dplyr package uses C++ code to evaluate. Let’s create another data frame with all numeric columns and run these examples.

# Create dataframe with numeric columnsmy_dataframe=data.frame(pages=c(32,45,NA,22,NA), chapters=c(NA,86,11,15,NA), price=c(144,553,321,567,NA))# Replace NA using coalesce() from dplyrlibrary("dplyr")my_dataframe <- mutate_all(my_dataframe, ~coalesce(.,0))# Replace NA using replace_na() from tidyrlibrary("dplyr")library("tidyr")my_dataframe <- mutate_all(my_dataframe, ~replace_na(.,0))# Replace NA using setnafill() from data.tablelibrary("data.table")my_dataframe <- setnafill(my_dataframe, fill=0)

All above examples yield the same below output.

# Output id pages chapters price1 11 32 0 1442 22 45 86 5533 33 0 11 3214 44 22 15 5675 0 0 0 0

Here, the coalesce() function is fromdplyrpackage. This returns the first non-missing value of its arguments.

6. Update NA with Zero By Specific Column Name

Here we use mutate() function with coalesce() from dplyr package. This updates NA values with zero on the id column. By using this on character columns you will get an error.

# Load dplyr librarylibrary("dplyr")#Replace NA with zero on specific numeric columnmy_dataframe <- my_dataframe %>% mutate(id = coalesce(id, 0))

7. Update NA with Zero on Multiple Columns by Name

Let’s use the same above approach but replace NA with zero on multiple columns by column name.

# Replace on multiple columnslibrary("dplyr")my_dataframe <- my_dataframe %>% mutate(id = coalesce(id, 0), pages = coalesce(pages, 0))

8. Replace NA with 0 on Column by Index

Use mutate_at() to specify the index number where you wanted to replace NA values with zero in R data frame.

# Load tidyr librarylibrary("tidyr")library("dplyr")my_dataframe <- my_dataframe %>% mutate_at(1, ~replace_na(.,0))print(my_dataframe)

Yields below output.

# Output id pages chapters price1 11 32 NA 1442 22 45 86 5533 33 NA 11 3214 44 22 15 5675 0 NA NA NA

9. Replace NA on Multiple Columns by Index

mutate_at() also takes vector with index numbers which is used to replace NA with 0 on multiple columns and replace_na() replaces all NA with 0.

# Replace NA on multiple columns by Indexlibrary("tidyr")library("dplyr")my_dataframe <- my_dataframe %>% mutate_at(c(1,3), ~replace_na(.,0))print(my_dataframe)

Yields below output.

# Output id pages chapters price1 11 32 0 1442 22 45 86 5533 33 NA 11 3214 44 22 15 5675 0 NA 0 NA

10. Replace Only on Numeric Columns

When you have data.frame with a mix of numeric and character columns, to update only numeric columns from NA with 0 use mutate_if() with is.numeric as a parameter.

# Replace only numeric columnslibrary("tidyr")library("dplyr")my_dataframe <- my_dataframe %>% mutate_if(is.numeric, ~replace_na(., 0))

11. Data with Factor Values

If you have data with numeric and characters most of the above examples work without issue. But, if you have factor values, first you need to convert them to a character before replacing NA with zero.

#Change factors to character typemy_dataframe[i] <- lapply(my_dataframe[i], as.character)# Replace NA with 0my_dataframe[is.na(my_dataframe)] <- 0 # Change character columns back to factorsmy_dataframe[i] <- lapply(my_dataframe[i], as.factor) 

12. Conclusion

In this article, I have explained several ways to replace NA values with zero (0) on numeric columns of R data frame. We can use replace() method in two ways. One is from imputeTS package and another way is we can use it directly.

Related Articles

• R – Replace Character in a String
• R – Replace Column Value with Another Column
• R dplyr::mutate() – Replace Column Values
• R – Replace String with Another String or Character
• R – Replace Values Based on Condition
• R – str_replace() to Replace Matched Patterns in a String.
• R – Replace Empty String with NA
• R – Replace Zero (0) with NA on Dataframe Column

References

1. replace() in R
2. imputeTS() package in R
3. NA