modify: In place modification of data tables

In skranz/dplyrExtras: extra functionality for dplyr like mutate_rows for mutation of a subset of rows

Description

If dt is a data table, then modify is essentially just a wrapper for data.table syntax that allows modification or creation of new columns. If dt is not a data table, it will by default be converted to a data table and then transformed and returned as the original data frame. Unlike mutate from dplyr, one can use the .SD argument in function calls, which can quite useful sometimes.

Usage

modify(.dt, .if, .by = NULL, ..., .envir = parent.frame(),
  .inplace = is.data.table(.dt), .as.data.table = is.data.table(.dt))

Arguments

.dt	a data.table
.if	optional a boolean conditions that specifies the rows that shall be modifed
.by	optional a vector of column names used for computations that are splitted by groups
...	formulas for columns that are modified or newly created
.envir	optional an environment in which the expressions shall be evaluated if variables are not found in .dt
.inplace	allows .dt inplace modification (TRUE if .dt is a data table)
.as.data.table	shall result be a data.table (only true if .dt is a data.table)

Examples

## Not run: 
  library(microbenchmark)

  K = 3
  n = 10
  dt = data.table(a= sample(1:3,n,replace=TRUE),
                   b= sample(1:100,n,replace=TRUE),
                   x=rnorm(n))
  df = as.data.frame(dt)
  # Set x to 100 where a==2
  modify(dt,a==2, y=x+100, z=y+K)
  
  modify(df,a==2,y=200,z=y*5+K)
  
  dt[,y:=x+2]
  
  # Set x to the mean value of b*100 in each group of a
  modify(dt,.by=c("a"), x=mean(b)*100)
  dt
  # Call with strings
  com = "x=200"
  s_modify(dt,"a==2", com)
  dt

 

## End(Not run)

skranz/dplyrExtras documentation built on May 20, 2020, 6:39 p.m.