dplyrcheat sheet covers every core function for filtering, selecting, mutating, grouping, summarising, and joining data in R — grouped by task, not alphabetically, with real output from the built-in mtcars dataset so you can check your own results against it.
Table of Contents
Key Takeaways
- dplyr is the tidyverse package for data wrangling in R — filtering, selecting, transforming, grouping, and joining data frames.
- The pipe (
%>%or base R's|>) chains functions left-to-right so you can read code as a sequence of steps instead of nested parentheses. filter(),select(),mutate(),group_by(), andsummarise()cover roughly 90% of everyday data wrangling tasks.- Joins (
left_join(),inner_join(),full_join(),anti_join()) combine two data frames on a shared key — this is the section most cheat sheets skip and the one you'll need most for real projects. - Every example below runs against R's built-in
mtcarsdataset, so you can paste the code into RStudio and get the exact same output shown here.
Install and Load dplyr
Install once, load every session:
install.packages("tidyverse") # installs dplyr + the rest of the tidyverse
# or, just dplyr on its own:
install.packages("dplyr")
library(dplyr)If you're new to R or RStudio itself, start here: Comprehensive Guide: How to Install RStudio.
The Pipe Operator: %>% and |>
The pipe passes the result on its left into the first argument of the function on its right. It turns nested function calls into a readable top-to-bottom sequence.
mtcars %>%
filter(mpg > 25) %>%
select(mpg, hp) %>%
arrange(desc(mpg))mtcars |>
filter(mpg > 25) |>
select(mpg, hp) |>
arrange(desc(mpg))|> is base R (4.1+, no package needed); %>% comes from magrittr/dplyr and has a few extra features (like placing the piped object anywhere with .). Either is fine — pick one and stay consistent within a script.
mtcars dataset (32 cars, no download needed) so you can run the exact code shown and get the exact output shown.Row Functions: Filter, Arrange, Slice, Distinct
Row functions return a subset or reordering of rows — the table stays the same shape, columns don't change.
| Function | What it does |
|---|---|
filter() | Keep rows matching a condition |
arrange() | Reorder rows by column value(s) |
slice() | Select rows by position |
distinct() | Remove or extract duplicate rows |
sample_n() / slice_sample() | Randomly sample rows |
top_n() / slice_max() | Select top/bottom N rows by a variable |
filter() — keep rows matching a condition
mtcars %>%
filter(mpg > 25) model mpg hp
Fiat 128 32.4 66
Honda Civic 30.4 52
Toyota Corolla 33.9 65
Fiat X1-9 27.3 66
Porsche 914-2 26.0 91
Lotus Europa 30.4 113Combine conditions with & (AND), | (OR), and %in% (value in a set):
mtcars %>% filter(mpg > 20 & cyl == 6) # AND
mtcars %>% filter(cyl == 4 | cyl == 8) # OR
mtcars %>% filter(cyl %in% c(4, 6)) # value in a setarrange() — reorder rows
mtcars %>%
arrange(desc(mpg)) %>%
select(mpg) %>%
head(3) model mpg
Toyota Corolla 33.9
Fiat 128 32.4
Lotus Europa 30.4Default is ascending; wrap the column in desc() for descending.
distinct() — unique rows
mtcars %>% distinct(cyl) cyl
1 6
2 4
3 8Column Functions: Select, Rename, Mutate, Transmute
Column functions change which columns exist or create new ones — row count stays the same.
| Function | What it does |
|---|---|
select() | Keep, drop, or reorder columns |
rename() | Rename a column, keep everything else |
relocate() | Move a column's position |
mutate() | Add or modify a column, keep the rest |
transmute() | Add a column, drop everything else |
select() — choose columns
mtcars %>% select(mpg, hp, cyl) # by name
mtcars %>% select(-cyl) # everything except cyl
mtcars %>% select(starts_with("m")) # by prefix
mtcars %>% select(mpg:hp) # range of columnsmutate() — add a new column
mtcars %>%
mutate(mpg_kpl = round(mpg * 0.425, 2)) %>%
select(mpg, mpg_kpl) %>%
head(3) model mpg mpg_kpl
Mazda RX4 21.0 8.92
Mazda RX4 Wag 21.0 8.92
Datsun 710 22.8 9.69Grouping and Summarising: group_by, summarise, count
This is the "split-apply-combine" pattern: split the data into groups, apply a summary function to each, combine the results into one table.
group_by() + summarise()
mtcars %>%
group_by(cyl) %>%
summarise(
avg_mpg = round(mean(mpg), 1),
avg_hp = round(mean(hp), 1),
n = n()
) cyl avg_mpg avg_hp n
1 4 26.7 82.6 11
2 6 19.7 122.3 7
3 8 15.1 209.2 14Always call ungroup() after you're done with grouped operations, or later steps in the pipeline will silently stay grouped.
count() — quick group counts
mtcars %>% count(cyl) cyl n
1 4 11
2 6 7
3 8 14Joining Two Data Frames
Joins combine two tables on a shared key column. This is the section most dplyr references skip — and the one you'll actually need once your data isn't already in one table.
cars <- tibble(
id = c(1, 2, 3, 4),
model = c("Civic", "Corolla", "Model 3", "Mustang")
)
prices <- tibble(
id = c(1, 2, 3, 5),
price = c(24000, 23500, 42000, 31000)
)| Join | Keeps | Code |
|---|---|---|
inner_join() | Only rows with a match in both tables | cars %>% inner_join(prices, by = "id") |
left_join() | All rows from the left table; unmatched right columns become NA | cars %>% left_join(prices, by = "id") |
right_join() | All rows from the right table | cars %>% right_join(prices, by = "id") |
full_join() | All rows from both tables | cars %>% full_join(prices, by = "id") |
anti_join() | Rows in the left table with NO match in the right | cars %>% anti_join(prices, by = "id") |
cars %>% left_join(prices, by = "id") id model price
1 1 Civic 24000
2 2 Corolla 23500
3 3 Model 3 42000
4 4 Mustang NAMustang (id 4) has no matching price, so left_join() keeps the row and fills price with NA. Swap to inner_join() and that row disappears entirely — that's the core distinction between the two.
Full dplyr Function Reference
Grouped by task for quick lookup. Copy the function name into RStudio's help (?function_name) for full argument details.
Row functions
| Function | Description |
|---|---|
filter() | Subset rows matching a condition |
arrange() | Reorder rows |
slice() | Select rows by position |
distinct() | Unique rows |
slice_sample() | Random row sample |
slice_max() / slice_min() | Top/bottom N rows by a variable |
row_number() | Row index within current order |
Column functions
| Function | Description |
|---|---|
select() | Keep/drop/reorder columns |
rename() | Rename a column |
relocate() | Move column position |
mutate() | Add/modify a column |
transmute() | Add a column, drop the rest |
across() | Apply one function across multiple columns |
if_else() / case_when() | Conditional value assignment |
coalesce() | First non-missing value across columns |
Grouping and summarising
| Function | Description |
|---|---|
group_by() | Group rows by one or more columns |
ungroup() | Remove grouping |
summarise() | Collapse each group to one summary row |
count() | Count rows per group |
n() | Count of rows in current group |
Joins
| Function | Description |
|---|---|
inner_join() | Matching rows only |
left_join() | All left rows |
right_join() | All right rows |
full_join() | All rows, both tables |
anti_join() | Left rows with no match in right |
bind_rows() / bind_cols() | Stack tables vertically/horizontally |
Frequently Asked Questions
Is there a downloadable PDF version of this dplyr cheat sheet?
Not yet as a standalone PDF — this page is a living reference we update as dplyr changes, which a static PDF can't do. You can bookmark or print this page directly from your browser (Ctrl/Cmd + P) for an offline copy.
What is dplyr used for in R?
dplyr provides a consistent set of functions — filter, select, mutate, group_by, summarise, and join — for wrangling data frames in R. It's part of the tidyverse.
How do I select specific columns with dplyr?
Use select(): df %>% select(col1, col2) keeps only those columns. Use select(-col1) to drop a column instead, or helpers like starts_with("x") to select by pattern.
What's the difference between dplyr and base R for data manipulation?
Base R uses bracket notation (df[df$mpg > 25, ]) and nested function calls. dplyr uses named verbs (filter(), select()) chained with the pipe, which is generally more readable for multi-step transformations and is the standard used across the tidyverse.
Is dplyr part of the tidyverse?
Yes — dplyr is one of the core tidyverse packages, alongside ggplot2, tidyr, and readr. Installing tidyverse installs dplyr automatically.
How do I install dplyr if install.packages() fails?
Confirm you have an active internet connection and try install.packages("dplyr") again. If it still fails, install the development version from GitHub: install.packages("pak"); pak::pak("tidyverse/dplyr").
What does the %>% pipe operator do in dplyr?
It passes the result on its left into the first argument of the function on its right, letting you chain multiple operations top-to-bottom instead of nesting function calls. Base R's built-in |> pipe works the same way without needing a package.
How do I join two data frames in dplyr?
Use left_join(), inner_join(), right_join(), or full_join() with a shared key column: df1 %>% left_join(df2, by = "id"). See the Joining Two Data Frames section above for what each one keeps.
Need this analysis done for your thesis or dissertation? I'll handle it in R, SPSS, or Minitab — with APA-formatted results delivered fast.
