Written and maintained by CASRAI Editorial Board
Published · Last updated
On this page
The rep function in R replicates elements of vectors, lists, factors, and date-time objects across controlled patterns and frequencies. Defined as a core primitive within base R, rep() allows investigators to construct balanced experimental designs, populate randomized trial allocations, replicate biological cohorts, and feed iterative Monte Carlo resampling loops without installing third-party dependencies.
What it is used for in research
In empirical and quantitative research, data structures frequently require balanced repetitions across conditions, subjects, treatments, and observation times. Researchers rely on the rep() function across four primary analytical workflows:
- Experimental Design and Factorial Blocks: Generating balanced allocations of experimental subjects across control and treatment conditions (for example, assigning 20 participants to Treatment A, 20 to Treatment B, and 20 to Placebo).
- Longitudinal and Repeated-Measures Data Structuring: Populating repeated subject identifiers, measurement waves, or temporal milestones across rectangular panel datasets before conducting statistical modeling in R.
- Monte Carlo Simulations and Bootstrapping: Generating baseline parameter vectors or repeating null-distribution indices across tens of thousands of resampled iterations.
- Preserving Complex Classes in Factor Matrices: Unlike rudimentary concatenation, generic
rep()preserves factor levels, date formats (Date,POSIXct), and S4 object structures, preventing subtle data loss during preprocessing.
How to install or access it
The rep() function is an integral component of R’s base package, maintained directly by The R Foundation and the R Core Team. Because base is pre-installed and loaded by default into every R environment, researchers do not need to call install.packages() or load external libraries.
To verify access or check the local R environment documentation, execute the following commands in the R terminal or RStudio console:
# Check current R release and core environment
sessionInfo()
# Open official base R documentation for rep
?base::rep
For custom research workflows requiring automated vector transformations, investigators often integrate rep() inside custom functions in R to standardize data preprocessing across multi-center trials.
Step-by-step instructions
The rep() generic function and its specialized primitives evaluate replication through four primary arguments: x (the input vector or object), times (the overall or per-element replication count), each (consecutive replications per element), and length.out (target output length). Follow these numbered steps to execute common research tasks:
1. Replicate an entire vector end-to-end using scalar times
To repeat a complete sequence of values sequentially, provide an integer scalar to the times argument:
# Replicate cohort sequence twice
cohorts <- c("Cohort_1", "Cohort_2", "Cohort_3")
rep(cohorts, times = 2)
# Output: [1] "Cohort_1" "Cohort_2" "Cohort_3" "Cohort_1" "Cohort_2" "Cohort_3"
2. Repeat each element consecutively using each
To repeat each individual element back-to-back before advancing to the subsequent element, specify the each argument:
# Repeat each treatment arm 3 times consecutively
treatments <- c("Placebo", "Drug_5mg", "Drug_10mg")
rep(treatments, each = 3)
# Output: [1] "Placebo" "Placebo" "Placebo" "Drug_5mg" "Drug_5mg" "Drug_5mg" "Drug_10mg" "Drug_10mg" "Drug_10mg"
3. Apply variable element-wise frequencies using vector times
When experimental groups require unequal sample sizes, pass a numeric vector to times matching the length of x:
# Assign unequal sample sizes across study arms
sample_groups <- c("Control", "Low_Dose", "High_Dose")
group_sizes <- c(10, 15, 20)
rep(sample_groups, times = group_sizes)
# Output: Returns a 45-element character vector with 10 Controls, 15 Low_Doses, 20 High_Doses
4. Enforce exact vector lengths with length.out
To restrict or extend an output sequence to an exact sample size regardless of whole multiples, set length.out:
# Constrain sequence to exactly 7 observation slots
measurements <- c("Pre", "Mid", "Post")
rep(measurements, length.out = 7)
# Output: [1] "Pre" "Mid" "Post" "Pre" "Mid" "Post" "Pre"
5. Combine each and times for nested factorial blocking
Researchers can combine each and a scalar times to generate multi-factorial blocks. In this combination, R first expands each element by each, and then replicates that complete series by times:
# Generate 2 complete blocks of 3 replicates per condition
rep(c("A", "B"), each = 3, times = 2)
# Output: [1] "A" "A" "A" "B" "B" "B" "A" "A" "A" "B" "B" "B"
6. Utilize high-performance primitives rep.int and rep_len for simulation
When executing iterative simulation algorithms (such as bootstrap resampling or MCMC runs with >100,000 cycles), bypass generic method dispatch by invoking the internal C-level primitives rep.int() and rep_len():
# High-speed replication of atomic integer vectors
resample_indices <- 1:500
sim_vector <- rep.int(resample_indices, times = 200)
# Pre-allocate exact length output vector
fixed_len <- rep_len(c(0L, 1L), length.out = 100000L)
Output and how to read it
The return value of rep() depends directly on the class and attributes of the input object x. Understanding the console output and object structure ensures accurate integration into research pipelines:
# Factor replication example
conditions <- factor(c("Low", "Medium", "High"), levels = c("Low", "Medium", "High"))
expanded_conditions <- rep(conditions, each = 2)
str(expanded_conditions)
# Output:
# Factor w/ 3 levels "Low","Medium",..: 1 1 2 2 3 3
# Date replication example
baseline_dates <- as.Date(c("2026-01-01", "2026-02-01"))
rep(baseline_dates, times = 2)
# Output:
# [1] "2026-01-01" "2026-02-01" "2026-01-01" "2026-02-01"
When inspecting output from rep():
- Atomic Vectors: Replicating numeric, integer, character, or logical vectors yields an atomic vector with the identical
typeof()andstorage.mode(). - Factors: Generic
rep()preserves the underlying factor levels and integer codes. However,rep.int()applied to factors in older R scripts can strip levels; use genericrep()when working with categorical variables. - Date-Time Objects: Classes like
Date,POSIXct, andPOSIXltpreserve their calendar attributes. R does not convert them into raw Unix timestamps. - Data Integrity: Ensuring exact alignment of replicated labels across observation tables protects research records from accidental row shifts, supporting compliance with ALCOA+ data integrity and facilitating reproducible outputs aligned with FAIR data principles.
Troubleshooting
When applying rep() across complex research scripts, analysts commonly encounter the following operational issues:
| Issue / Error Message | Root Cause | Recommended Solution |
|---|---|---|
Error in rep(...) : invalid 'times' argument |
Supplying a vector to times whose length does not equal length(x), or supplying negative or non-finite values. |
Verify that length(times) == length(x) when using vector-valued times, or pass a single non-negative integer scalar. |
Combining vector times with each |
Base R does not permit specifying both an element-wise vector times and an each argument simultaneously. |
Separate the transformation into two explicit steps, or use base::gl() to generate structured experimental factor layouts. |
Unexpected coercion with NA |
Using unadorned NA defaults to logical mode (NA), which may inadvertently coerce numeric columns when merged. |
Supply typed missing values (such as NA_real_, NA_integer_, or NA_character_) when pre-allocating missing rows. |
| Stripped attributes on custom objects | Invoking rep.int() on complex S3/S4 objects or factors strips custom non-vector attributes for speed. |
Use the generic rep() function whenever object class dispatch and attribute preservation are required. |
Free alternatives
While base::rep() is standard for vector replication, statistical programming in R and related open-source ecosystems offers several specialized alternatives depending on data type and computational context:
| Function / Tool | Package / Ecosystem | Primary Use Case | Key Advantage | Trade-Off / Limitation |
|---|---|---|---|---|
base::rep() |
Base R | General vector, list, and factor replication | Pre-installed; preserves classes such as Date and factor |
Dispatches S3 methods, introducing slight overhead in tight loops |
base::rep.int() |
Base R | High-speed atomic vector replication | Internal C primitive; optimal speed in simulation loops | Does not support each; strips non-core attributes |
base::rep_len() |
Base R | Replication constrained to exact output length | Internal C primitive; memory-efficient length pre-allocation | Only accepts length.out; no per-element frequency control |
base::gl() |
Base R | Generating factor levels for experimental designs | Creates balanced factor variables with labels and orders in one call | Restricted to integer factor levels with equal or structured replications |
base::seq() |
Base R | Generating regular arithmetic sequences | Calculates incremental numeric steps across ranges (by, from, to) |
Generates numerical progressions rather than repeating static elements |
vctrs::vec_rep() |
Tidyverse (vctrs) |
Type-stable replication of data frames and vectors | Strict type and size validation; replicates entire data frame rows | Requires external package dependency (vctrs) |
numpy.repeat() |
Python (numpy) |
Array element repetition in scientific Python | C-level performance across multidimensional matrix axes | Requires Python runtime; zero-indexed coordinate system |
FAQ
What is the difference between times and each in the R rep function?
The times argument repeats the entire vector sequentially (or applies a vector of frequencies per element), whereas each repeats every individual element consecutively before moving to the next element. For example, rep(1:2, times = 2) produces 1, 2, 1, 2, while rep(1:2, each = 2) produces 1, 1, 2, 2.
When should researchers use rep.int() instead of rep()?
Researchers should use rep.int() when conducting computationally intensive operations—such as Monte Carlo simulations, permutation tests, or bootstrap resampling exceeding 100,000 iterations—on atomic vectors. Because rep.int() is an internal C primitive that bypasses method dispatch, it executes significantly faster than generic rep().
How does rep() handle factors and date objects in R?
The generic rep() function dispatches specialized methods for factors and date-time classes (Date, POSIXct, POSIXlt). It preserves factor levels and calendar attributes intact. In contrast, using rep.int() on factors drops factor levels and returns raw integer codes.
Can you use rep() to replicate rows in an R data frame?
Yes. Although rep() operates directly on vectors and lists, you can replicate data frame rows by passing repeated index numbers to the row-indexing bracket: df[rep(1:nrow(df), each = 2), ]. Alternatively, the tidyverse function vctrs::vec_rep() replicates whole data frame rows natively.
Why does rep(1:3, times = c(2, 3)) produce an error?
When times is supplied as a vector, R requires its length to match exactly the length of the input vector x. Because 1:3 contains 3 elements and c(2, 3) contains only 2 elements, R throws Error in rep(1:3, times = c(2, 3)) : invalid 'times' argument.
Sources
- The R Foundation / R Core Team. “Replicate Elements of Vectors and Lists (
base::rep)”. R Documentation, R 4.4.x / Devel. Accessed November 10, 2026. https://stat.ethz.ch/R-manual/R-devel/library/base/html/rep.html - Comprehensive R Archive Network (CRAN), R Core Team. “An Introduction to R: Notes on R: A Programming Environment for Data Analysis and Graphics”, Section 2.3: “Generating regular sequences”. Accessed November 10, 2026. https://cran.r-project.org/doc/manuals/r-release/R-intro.html#Generating-regular-sequences
- Comprehensive R Archive Network (CRAN), R Core Team. “R Language Definition: Objects, Evaluation and Primitives”. Accessed November 10, 2026. https://cran.r-project.org/doc/manuals/r-release/R-lang.html







