mv: Functions to manipulate variables' folder structure

Description

Variables in Crunch datasets are organized into folders, like in a file system. These functions allow you to create new folders and move objects into folders. Their names, mv and mkdir, suggest their Unix file utility inspiration.

Usage

mv(dataset, variables, path)
mkdir(dataset, path, variables = NULL)

Arguments

dataset

A Crunch dataset

variables

A Variable, selection of variables from dataset, or any other object that can be moved to a folder. For mkdir, variables is optional.

path

A character "path" to the folder: either a vector of nested folder names or a single string with nested folders separated by a delimiter ("/" default, configurable via options(crunch.delimiter))

Value

dataset, with the folder at path guaranteed to be created and variables, if specified, moved into it.

Details

The functions have some differences from the strict behavior of their Unix ancestors. For one, they work recursively, without additional arguments: mkdir will make every directory necessary to construct the requested path, even if all parent directories didn't already exist; and mv doesn't require that the directory to move to already exist. mkdir also takes an optional variables argument, which allows you to say "make this directory and put these variables in it" in one command. As a result of these behaviors, mv and mkdir here are essentially the same, just with arguments in reverse order: mv is move "variables" to "path", while mkdir is create "path" and put "variables" in it.

Both functions take "dataset" as the first argument, so they naturally support pipelining (as with the %>% operator).

Examples

Run this code

# NOT RUN {
ds <- loadDataset("Example survey")
ds <- mv(ds, c("gender", "age", "educ"), "Demographics")
ds <- mkdir(ds, "Key Performance Indicators/Brand X")
# These can also be chained together
require(magrittr)
ds <- ds %>%
    mv(c("aware_x", "nps_x"), "Key Performance Indicators/Brand X") %>%
    mv(c("aware_y", "nps_y"), "Key Performance Indicators/Brand Y")
# }