In WinVector/rquery: Relational Query Generator for Data Manipulation at Scale

library("dplyr")
library("rqdatatable")
library("microbenchmark")

batting <- Lahman::Batting


# dplyr pipeline example from R for Data Science.
fn_dplyr <- function(batting) {
  # http://r4ds.had.co.nz/transform.html
  batting %>% 
    group_by(playerID) %>% 
    summarise(
      ba = sum(H, na.rm = TRUE) / sum(AB, na.rm = TRUE),
      ab = sum(AB, na.rm = TRUE)
    ) %>%
    filter(ab > 100) 
}

system.time(
  batters <- fn_dplyr(batting)
)
batters <- arrange(batters, playerID)

# translation of above example into an rquery pipeline
fn_rquery <- function(batting) {
  rq_pipeline <- local_td(batting) %.>%
    project_nse(., groupby = "playerID",
                ba = sum(H) / sum(AB),
                ab = sum(AB)) %>%
    select_rows_nse(., ab > 100) %>%
    orderby(., "playerID")
  ex_data_table(rq_pipeline)
}

system.time({
  batters2 <- fn_rquery(batting)
})


assertthat::are_equal(as.data.frame(batters),
                      as.data.frame(batters2))

microbenchmark(
  dplyr = nrow(fn_dplyr(batting)),
  rquery = nrow(fn_rquery(batting))
)