This function gives you a Dataset object, which refers to a dataset hosted on the Crunch platform. With this Dataset, you can perform lots of data cleaning and analysis as if the dataset were fully resident on your computer, without having to pull data locally.

  kind = c("active", "all", "archived"),
  project = NULL,
  refresh = FALSE



character, the name or path to a Crunch dataset to load, or a dataset URL. If dataset is a path to a dataset in a project, the path will be be parsed and walked, relative to project if specified, and the function will look for the dataset inside that project. If no path is specified and no project provided, the function will call a search API to do an exact string match on dataset names.


character specifying whether to look in active, archived, or all datasets. Default is "active", i.e. non-archived.


ProjectFolder entity, character name (path) to a project, or NULL, the default. If a Project entity or reference is supplied, either here or as a path in dataset, the dataset lookup will be limited to that project only.


logical: should the function check the Crunch API for new datasets? Default is FALSE.


An object of class CrunchDataset.


You can specify a dataset to load by its human-friendly "name", possibly also by indicating a project (folder) to find it in. This makes code more readable, but it does mean that if the dataset is renamed or moved to a different folder, your code may no longer work. The fastest, most reliable way to use loadDataset() is to provide a URL to the dataset--the dataset's URL will never change.

See also

See cd() for details of parsing and walking dataset folder/project paths.


if (FALSE) {
ds <- loadDatasets("A special dataset")
ds2 <- loadDatasets("~/My dataset")
ds3 <- loadDataset("My dataset", project = "~") # Same as ds2
ds4 <- loadDataset("")