Alternate assignment (multiple and/or collect results from dplyr)

These alternate assignment operators can be used to perform multiple assignment (also known as destructuring assignment). These are imported from the {zeallot} package (see the corresponding help page at zeallot::operator for complete description). They also performs a dplyr::collect() allowing to get results from dplyr extensions like {dtplyr} for data.tables, or {dbplyr} for databases. Finally these two assignment operators also make sure that the preferred data frame object is returned by using default_dtx().

value %->% x

x %<-% value

# Default S3 method
collect(x, ...)

Arguments

value

The object to be assigned.

x

A name, or a name structure for multiple (deconstructing) assignment, or any object that does not have a specific [dplyr::collect[]) method for collect.default().

...

further arguments passed to the method (not used for the default one)

Value

These operators invisibly return value. collect.default() simply return x.

Details

These assignation operator are overloaded to get interesting properties in the context of {tidyverse} pipelines and to make sure to always return our preferred data frame object (data.frame, data.table, or tibble). Thus, before being assigned, value is modified by calling dplyr::collect() on it and by applying default_dtx().

Examples

# The alternate assignment operator performs three steps:
# 1) Collect results from dbplyr or dtplyr
library(dplyr)
#> 
#> Attaching package: ‘dplyr’
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, lag
#> The following objects are masked from ‘package:base’:
#> 
#>     intersect, setdiff, setequal, union
library(data.table)
#> 
#> Attaching package: ‘data.table’
#> The following objects are masked from ‘package:dplyr’:
#> 
#>     between, first, last
library(dtplyr)
library(svBase)
dtt <- data.table(x = 1:5, y = rnorm(5))
dtt |>
  mutate(x2 = x^2) |>
  select(x2, y) ->
  res

print(res)
#>       x2            y
#>    <num>        <num>
#> 1:     1 -2.437263611
#> 2:     4 -0.005571287
#> 3:     9  0.621552721
#> 4:    16  1.148411606
#> 5:    25 -1.821817661
class(res) # This is a data frame
#> [1] "data.table" "data.frame"

dtt |>
  lazy_dt() |>
  mutate(x2 = x^2) |>
  select(x2, y) ->
  res

print(res)
#> Source: local data table [5 x 2]
#> Call:   setcolorder(copy(`_DT1`)[, `:=`(x2 = x^2)][, `:=`("x", NULL)], 
#>     c("x2", "y"))
#> 
#>      x2        y
#>   <dbl>    <dbl>
#> 1     1 -2.44   
#> 2     4 -0.00557
#> 3     9  0.622  
#> 4    16  1.15   
#> 5    25 -1.82   
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results
class(res) # This is NOT a data frame
#> [1] "dtplyr_step_call" "dtplyr_step"     

# Same pipeline, but assigning with %->%
dtt |>
  lazy_dt() |>
  mutate(x2 = x^2) |>
  select(x2, y) %->%
  res

print(res)
#> # A data.trame: [5 × 2]
#>      x2        y
#>   <dbl>    <dbl>
#> 1     1 -2.44   
#> 2     4 -0.00557
#> 3     9  0.622  
#> 4    16  1.15   
#> 5    25 -1.82   
class(res) # res is the preferred data frame (data.table by default)
#> [1] "data.trame" "data.frame"

# 2) Convert data frame in the chosen format using default_dtx()
dtf <- data.frame(x = 1:5, y = rnorm(5))
class(dtf)
#> [1] "data.frame"
res %<-% dtf
class(res) # A data.table by default
#> [1] "data.trame" "data.frame"
# but it can be changed with options("SciViews.as_dtx)

# 3) If the zeallot syntax is used, make multiple assignment
c(X, Y) %<-% dtf # Variables of dtf assigned to different names
X
#> [1] 1 2 3 4 5
Y
#> [1] -0.2473253 -0.2441996 -0.2827054 -0.5536994  0.6289820

# The %->% is meant to be used in pipelines, otherwise it does the same