Repository files navigation

dtplyrprovides adata.table backend for dplyr. Thegoal of dtplyr is to allow you to write dplyr code that is automaticallytranslated to the equivalent, but usually much faster, data.table code.

Seevignette("translation") for details of the current translations,andtable.express andrqdatatable for relatedwork.

You can install from CRAN with:

install.packages("dtplyr")

Or try the development version from GitHub with:

# install.packages("pak")pak::pak("tidyverse/dtplyr")

To use dtplyr, you must at least load dtplyr and dplyr. You may alsowant to loaddata.table so you can access theother goodies that it provides:

library(data.table)library(dtplyr)library(dplyr,warn.conflicts=FALSE)

Then uselazy_dt() to create a “lazy” data table that tracks theoperations performed on it.

mtcars2<- lazy_dt(mtcars)

You can preview the transformation (including the generated data.tablecode) by printing the result:

mtcars2 %>%  filter(wt<5) %>%  mutate(l100k=235.21/mpg) %>%# liters / 100 km  group_by(cyl) %>%  summarise(l100k= mean(l100k))#> Source: local data table [3 x 2]#> Call:   `_DT1`[wt < 5][, `:=`(l100k = 235.21/mpg)][, .(l100k = mean(l100k)),#>     keyby = .(cyl)]#>#>     cyl l100k#>   <dbl> <dbl>#> 1     4  9.05#> 2     6 12.0#> 3     8 14.9#>#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

But generally you should reserve this only for debugging, and useas.data.table(),as.data.frame(), oras_tibble() to indicate thatyou’re done with the transformation and want to access the results:

mtcars2 %>%  filter(wt<5) %>%  mutate(l100k=235.21/mpg) %>%# liters / 100 km  group_by(cyl) %>%  summarise(l100k= mean(l100k)) %>%  as_tibble()#> # A tibble: 3 × 2#>     cyl l100k#>   <dbl> <dbl>#> 1     4  9.05#> 2     6 12.0#> 3     8 14.9

There are two primary reasons that dtplyr will always be somewhat slowerthan data.table:

• Each dplyr verb must do some work to convert dplyr syntax todata.table syntax. This takes time proportional to the complexity ofthe input code, not the inputdata, so should be a negligibleoverhead for large datasets.Initialbenchmarkssuggest that the overhead should be under 1ms per dplyr call.

• To match dplyr semantics,mutate() does not modify in place bydefault. This means that most expressions involvingmutate() mustmake a copy that would not be necessary if you were using data.tabledirectly. (You can opt out of this behaviour inlazy_dt() withimmutable = FALSE).

Please note that the dtplyr project is released with aContributor Codeof Conduct. Bycontributing to this project, you agree to abide by its terms.