dplyr dense_rank() in R: Rank With Ties Sharing Same Position

The dense_rank() function in dplyr ranks values where tied entries share the same rank, and the NEXT rank does NOT skip. The output covers a contiguous range 1, 2, 3, ... up to the number of distinct values.

By Selva Prabhakaran · Published May 12, 2026 · Last updated May 12, 2026

⚡ Quick Answer

dense_rank(c(10, 20, 20, 30))      # 1, 2, 2, 3 (no gap)
min_rank(c(10, 20, 20, 30))        # 1, 2, 2, 4 (gap)
dense_rank(desc(x))                 # rank descending
df |> mutate(level = dense_rank(score))
df |> group_by(g) |> mutate(level = dense_rank(score))
length(unique(x))                   # max rank from dense_rank

Need explanation? Read on for examples and pitfalls.

📊 Which ranking function?

What dense_rank() does in one sentence

dense_rank(x) returns the rank of each element where TIED values share the same rank and subsequent ranks DO NOT SKIP. For c(10, 20, 20, 30), the result is c(1, 2, 2, 3): distinct values get distinct, contiguous ranks.

This is the "dense" or "no-gap" ranking style. The maximum rank equals length(unique(x)).

Syntax

dense_rank(x). Use desc(x) for descending. NAs stay as NA in the output.

Run live

Run live, no install needed. Every R block on this page runs in your browser. Click Run, edit the code, re-run instantly. No setup.

RDense rank ascending

library(dplyr) x <- c(10, 20, 20, 30, 15) dense_rank(x) #> [1] 1 3 3 4 2

Tip

dense_rank is the right tool when you want to bucket values into ordered categories. Distinct values get distinct ranks; ties get the same rank. The result is a small contiguous integer range, perfect for category mapping.

Five common patterns

1. Rank with ties sharing position

RThree distinct values, two of them tied

dense_rank(c(50, 80, 80, 60, 90)) #> [1] 1 3 3 2 4

The two 80s share rank 3. The 90 (next distinct value) is rank 4, not 5.

2. Descending dense rank

RHighest value ranks 1

dense_rank(desc(c(50, 80, 80, 60, 90))) #> [1] 4 2 2 3 1

3. Count distinct values via dense_rank

RMax dense rank = number of distinct values

x <- c(10, 20, 20, 30, 15, 30) max(dense_rank(x)) #> [1] 4 length(unique(x)) #> [1] 4

A useful identity: max(dense_rank(x)) always equals length(unique(x)) (excluding NAs).

4. Per-group dense rank

RRestart ranks at each group

df_g <- data.frame( team = c("A","A","B","B","B"), pts = c(10, 20, 5, 15, 15) ) df_g |> group_by(team) |> mutate(level = dense_rank(desc(pts))) #> # A tibble: 5 x 3 #> team pts level #> A 10 2 #> A 20 1 #> B 5 2 #> B 15 1 #> B 15 1

5. Bucket continuous values into ordered levels

RCategorical encoding via dense_rank

df <- data.frame(price = c(10, 50, 100, 100, 25)) df |> mutate(price_level = dense_rank(price)) #> price price_level #> 1 10 1 #> 2 50 3 #> 3 100 4 #> 4 100 4 #> 5 25 2

price_level is now a small integer range (1 to 4) suitable as a categorical input.

Key Insight

dense_rank is the only rank function whose output forms a contiguous integer range. min_rank and row_number can leave gaps or break ties. dense_rank always covers 1..k where k is the number of distinct values. Useful for mapping continuous values to ordinal categories.

dense_rank() vs min_rank() vs row_number() vs ntile()

Four ranking-style functions in dplyr.

Function	Output for c(10, 20, 20, 30)	Best for
`dense_rank()`	1, 2, 2, 3	Categorical encoding; no gaps
`min_rank()`	1, 2, 2, 4	Standard rank with gaps
`row_number()`	1, 2, 3, 4	Unique sequential IDs
`ntile(x, 3)`	1, 2, 2, 3	Bin into n equal-sized groups

When to use which:

dense_rank for "ordinal level" or "tier" assignment.
min_rank for sports / standings.
row_number for unique row IDs.
ntile for binning into quantiles.

A practical workflow

The "tier assignment" pattern is the dense_rank sweet spot.

RTier assignment via dense_rank

products |> mutate(tier = dense_rank(desc(rating)))

A 5-star product and a 4.9-star product get tier 1 and 2; ties on the same rating share a tier. The maximum tier is the number of distinct ratings.

For per-category tiers:

RPer-category tier assignment

products |> group_by(category) |> mutate(tier = dense_rank(desc(rating))) |> ungroup()

Each category has its own contiguous tier range.

Common pitfalls

Pitfall 1: confusing dense_rank with min_rank. min_rank(c(10, 20, 20, 30)) is c(1, 2, 2, 4) (gap). dense_rank(c(10, 20, 20, 30)) is c(1, 2, 2, 3) (no gap). Pick based on whether downstream code expects gaps.

Pitfall 2: NA propagation. NAs stay as NA in the rank vector. Filter NAs first or set na.last = NA (not directly supported; use dense_rank(if_else(is.na(x), -Inf, x)) as a workaround).

Warning

dense_rank returns the SAME rank for ALL ties. If you need unique ranks even when values are equal, use row_number instead.

Try it yourself

Try it: Encode the unique cyl values in mtcars as small integers 1, 2, 3 (lowest cyl = 1). Save to ex_cyl_level.

RYour turn: encode cyl as ordinal

ex_cyl_level <- mtcars |> # your code here head(ex_cyl_level) #> Expected: original rows + cyl_level column with 1, 2, or 3

Click to reveal solution

RSolution

ex_cyl_level <- mtcars |> mutate(cyl_level = dense_rank(cyl)) head(ex_cyl_level[, c("cyl", "cyl_level")]) #> cyl cyl_level #> Mazda RX4 6 2 #> Mazda RX4 Wag 6 2 #> Datsun 710 4 1 #> Hornet 4 Drive 6 2 #> Hornet Sportabout 8 3 #> Valiant 6 2

Explanation: dense_rank(cyl) maps cyl 4 -> 1, cyl 6 -> 2, cyl 8 -> 3. Each unique cyl gets a contiguous integer level.

After mastering dense_rank, look at:

min_rank(): gaps after ties
row_number(): unique sequential
percent_rank() / cume_dist(): percentile rankings
ntile(): bin into n equal-sized groups
factor(): convert dense ranks to factor levels
forcats::fct_inseq(): factor in numerical order

For converting dense ranks to factors with sensible levels, factor(dense_rank(x)) works.

FAQ

What does dense_rank do in dplyr?

dense_rank(x) ranks values where tied entries share the same rank, and subsequent ranks DO NOT skip. The output is a contiguous integer range 1..k.

What is the difference between dense_rank and min_rank?

dense_rank produces no gaps after ties (1, 2, 2, 3). min_rank leaves gaps (1, 2, 2, 4). dense_rank is useful when counting distinct rank levels matters.

How do I count the number of distinct ranks?

max(dense_rank(x)) returns the count of distinct values (ignoring NAs). Equivalent to length(unique(x[!is.na(x)])).

How do I rank descending with dense_rank?

dense_rank(desc(x)). Highest value gets rank 1.

Can dense_rank be used as a categorical encoder?

Yes. dense_rank(x) maps distinct values of x to a contiguous integer range, which is the ordinal-encoding pattern.

Navigate

Tidyverse packages

Deep dives

Wrangling & EDA

Statistics

Machine Learning

Time Series

By Industry

Reporting & Apps

Levels

dplyr dense_rank() in R: Rank With Ties Sharing Same Position

What dense_rank() does in one sentence

Syntax

Five common patterns

1. Rank with ties sharing position

2. Descending dense rank

3. Count distinct values via dense_rank

4. Per-group dense rank

5. Bucket continuous values into ordered levels

dense_rank() vs min_rank() vs row_number() vs ntile()

A practical workflow

Common pitfalls

Try it yourself

FAQ

Navigate

Tidyverse packages

Deep dives

Wrangling & EDA

Statistics

Machine Learning

Time Series

By Industry

Reporting & Apps

Levels

dplyr dense_rank() in R: Rank With Ties Sharing Same Position

What dense_rank() does in one sentence

Syntax

Five common patterns

1. Rank with ties sharing position

2. Descending dense rank

3. Count distinct values via dense_rank

4. Per-group dense rank

5. Bucket continuous values into ordered levels

dense_rank() vs min_rank() vs row_number() vs ntile()

A practical workflow

Common pitfalls

Try it yourself

Related dplyr functions

FAQ