Package About Data Manipulation in Pure Base R

A lightweight, dependency-free data engine for R that provides a grammar for tabular and time-series manipulation. Built entirely on Base R, 'm61r' offers a fluent, chainable API inspired by modern data tools while prioritizing memory efficiency and speed. It includes optimized versions of common data verbs such as filtering, mutation, grouped aggregation, and approximate temporal joins, making it an ideal choice for environments where external dependencies are restricted or where performance in pure R is required.

m61r: The Pure Base R Data Engine 🌌

m61r provides dependency-free tabular and time-series manipulation for R. It is designed for the Base R purist who demands modern syntax and high performance.

Why m61r?

• Zero Dependencies: No dplyr, no tidyverse. It runs on any vanilla R installation.

• Unified Logic: Treat tabular and time series as one.

• High Performance: Optimized Base R verbs for tabular accounting.

• Fluent API: A clean, object-oriented API inspired by the best modern data tools

• Transparent: Written in 100% Base R

Quick Start

1. Tabular Manipulation

library(m61r)

# Initialize with the built-in CO2 dataset

co2 <- m61r(CO2)

co2$filter(~uptake > 30)$
    select(~c(Plant, Type, uptake))$
    group_by(~Type)$
    summarise(mean_uptake = ~mean(uptake))

print(co2[])

1. Temporal Data Manipulation

ts_data <- data.frame(
  time = as.POSIXct("2025-01-01 00:00:00") + runif(50, 0, 86400),
  consumption = rnorm(50, 500, 100)
)

p_agg <- m61r(ts_data)

p_agg$mutate(hour_bin = ~format(time, "%Y-%m-%d %H:00"))
p_agg$group_by(~hour_bin)
p_agg$summarise(
  n_obs = ~length(consumption),
  avg_load = ~mean(consumption)
)
p_agg$head(5)

Benchmarks

1. filter, mutate, head

library(microbenchmark)
library(m61r)

set.seed(42)
n <- 1e6
df_large <- data.frame(
  id = 1:n,
  group = sample(letters, n, replace = TRUE),
  val = rnorm(n)
)

p <- m61r(df_large)

results <- microbenchmark(
  # Approche Base R classique
  base_r = {
    tmp <- df_large[df_large$val > 0, ]
    tmp$new_val <- tmp$val * 2
    head(tmp)
  },
  
  # Approche m61r
  m61r_pipe = {    
    p$filter(~val > 0)
    p$mutate(new_val = ~val * 2)
    p$head()
  },
  times = 50 
)

print(results)

Unit: milliseconds

1. group_by, summarise

results_group <- microbenchmark(
  # Base R avec aggregate
  base_aggregate = {
    aggregate(val ~ group, data = df_large, FUN = mean)
  },
  
  # m61r
  m61r_grouped = {
    p$group_by(~group)
    p$summarise(avg = ~mean(val))
    p[]
  },
  times = 20
)

print(results_group)

Unit: milliseconds

Installation

Installation from CRAN

install.packages("m61r")

Installation from GITHUB in Base R

setwd("~")

download.file("https://github.com/pv71u98h1/m61r/archive/0.1..zip",destfile="m61r-0.1.0.zip")
unzip("m61r-0.1.0.zip")

# install
install.packages("m61r", repos=NULL, type='source')

unlink(file.path("~","m61r-0.1.0"),recursive=TRUE)
file.remove(file.path("~","m61r-0.1.0.zip"))