Data transformation with dplyr

Master data transformation with this comprehensive cheatsheet covering all essential dplyr functions. Learn how to efficiently filter rows, select columns, create new variables, and summarize data using the grammar of data manipulation.

dplyr functions work with pipes and expect tidy data. In tidy data:

• Each variable is in its own column
• Each observation, or case, is in its own row

Pipes: x |> f(y) becomes f(x, y)

1

library(dplyr)

Summarize cases

Create summary statistics tables:

• summarize() - Compute table of summaries
• count() - Count rows in each group
• tally(), add_count(), add_tally() - Additional counting helpers

Group cases

• group_by() - Create grouped copies for separate manipulation
• .by argument - Alternative syntax within functions
• rowwise() - Group data into individual rows
• ungroup() - Return ungrouped copy

Manipulate cases

Extract cases

• filter() - Extract rows meeting logical criteria
• distinct() - Remove duplicate rows
• slice() - Select rows by position
• slice_sample() - Randomly select rows
• slice_min() / slice_max() - Select lowest/highest values
• slice_head() / slice_tail() - Select first/last rows

Arrange cases

• arrange() - Order rows; use desc() for reverse order

Add cases

• add_row() - Add one or more rows

Manipulate variables

Extract variables

• pull() - Extract column as vector
• select() - Extract columns as table
• relocate() - Move columns to new positions

Select helpers

Use with select(): contains(), ends_with(), starts_with(), matches(), everything(), num_range(), : range notation, all_of(), any_of(), negation with !

Multiple variables

• across() - Summarize/mutate multiple columns similarly
• c_across() - Compute across columns in row-wise data

Make new variables

• mutate() - Compute new columns
• rename() - Rename columns
• rename_with() - Rename using a function

Vectorized functions

For use with mutate():

Offset: lag(), lead()

Cumulative aggregate: cumall(), cumany(), cummax(), cummean(), cummin(), cumprod(), cumsum()

Ranking: cume_dist(), dense_rank(), min_rank(), ntile(), percent_rank(), row_number()

Math: Arithmetic operators, logarithm functions, comparison operators, between(), near()

Miscellaneous: case_when(), coalesce(), if_else(), na_if(), pmax(), pmin()

Summary functions

For use with summarize():

Count: n(), n_distinct(), sum(!is.na())

Position: mean(), median()

Logical: mean() for proportions of TRUE, sum() for counts of TRUE

Order: first(), last(), nth()

Rank: quantile(), min(), max()

Spread: IQR(), mad(), sd(), var()

Combine tables

Combine variables

• bind_cols() - Place tables side by side

Combine cases

• bind_rows() - Stack tables vertically

Mutating joins

• left_join() - Include all rows from left table
• right_join() - Include all rows from right table
• inner_join() - Include only matching rows
• full_join() - Include all rows from both tables

Filtering joins

• semi_join() - Rows of x with matches in y
• anti_join() - Rows of x without matches in y

Nest join

• nest_join() - Inner join with nested results

Use by = join_by() for specifying column matches.

Set operations

intersect(), setdiff(), union(), union_all(), setequal()