class datalad.api.Dataset(*args, **kwargs)[source]

Representation of a DataLad dataset/repository

This is the core data type of DataLad: a representation of a dataset. At its core, datasets are (git-annex enabled) Git repositories. This class provides all operations that can be performed on a dataset.

Creating a dataset instance is cheap, all actual operations are delayed until they are actually needed. Creating multiple Dataset class instances for the same Dataset location will automatically yield references to the same object.

A dataset instance comprises of two major components: a repo attribute, and a config attribute. The former offers access to low-level functionality of the Git or git-annex repository. The latter gives access to a dataset’s configuration manager.

Most functionality is available via methods of this class, but also as stand-alone functions with the same name in datalad.api.


path (str or Path) – Path to the dataset location. This location may or may not exist yet.



type path:

add_archive_content(*[, dataset, annex, ...])

Add content of an archive under git annex control.

add_readme(*[, dataset, existing])

Add basic information about DataLad datasets to a README file

addurls(urlformat, filenameformat, *[, ...])

Create and update a dataset from a list of URLs.

clean(*[, what, dry_run, recursive, ...])

Clean up after DataLad (possible temporary files etc.)

clone([path, git_clone_opts, dataset, ...])

Obtain a dataset (copy) from a URL or local directory


Perform operations which would close any possible process using this Dataset

configuration([spec, scope, dataset, ...])

Get and set dataset, dataset-clone-local, or global configuration

copy_file(*[, dataset, recursive, ...])

Copy files and their availability metadata from one dataset to another.

create([initopts, force, description, ...])

Create a new dataset from scratch.

create_sibling(*[, name, target_dir, ...])

Create a dataset sibling on a UNIX-like Shell (local or SSH)-accessible machine

create_sibling_gin(*[, dataset, recursive, ...])

Create a dataset sibling on a GIN site (with content hosting)

create_sibling_gitea(*[, dataset, ...])

Create a dataset sibling on a Gitea site

create_sibling_github(*[, dataset, ...])

Create dataset sibling on (or an enterprise deployment).

create_sibling_gitlab(*[, site, project, ...])

Create dataset sibling at a GitLab site

create_sibling_gogs(*[, api, dataset, ...])

Create a dataset sibling on a GOGS site

create_sibling_ria(name, *[, dataset, ...])

Creates a sibling to a dataset in a RIA store

diff(*[, fr, to, dataset, annex, untracked, ...])

Report differences between two states of a dataset (hierarchy)

download_url(*[, dataset, path, overwrite, ...])

Download content

drop(*[, what, reckless, dataset, ...])

Drop content of individual files or entire (sub)datasets

export_archive(*[, dataset, archivetype, ...])

Export the content of a dataset as a TAR/ZIP archive.

export_archive_ora([opts, dataset, remote, ...])

Export an archive of a local annex object store for the ORA remote.

export_to_figshare(*[, dataset, ...])

Export the content of a dataset as a ZIP archive to figshare

foreach_dataset(*[, cmd_type, dataset, ...])

Run a command or Python code on the dataset and/or each of its sub-datasets.

get(*[, source, dataset, recursive, ...])

Get any dataset content (files/directories/subdatasets).

get_superdataset([datalad_only, topmost, ...])

Get the dataset's superdataset

install(*[, source, dataset, get_data, ...])

Install one or many datasets from remote URL(s) or local PATH source(s).


Returns whether a dataset is installed.

no_annex(pattern[, ref_dir, makedirs])

Configure a dataset to never put some content into the dataset's annex

push(*[, dataset, to, since, data, force, ...])

Push a dataset to a known sibling.


Something that can be used to checkout a particular state (tag, commit) to "undo" a change or switch to a otherwise desired previous state.

remove(*[, dataset, drop, reckless, ...])

Remove components from datasets

rerun(*[, since, dataset, branch, message, ...])

Re-execute previous datalad run commands.

run(*[, dataset, inputs, outputs, expand, ...])

Run an arbitrary shell command and record its impact on a dataset.

run_procedure(*[, dataset, discover, help_proc])

Run prepared procedures (DataLad scripts) on a dataset

save(*[, message, dataset, version_tag, ...])

Save the current state of a dataset

siblings(*[, dataset, name, url, pushurl, ...])

Manage sibling configuration

status(*[, dataset, annex, untracked, ...])

Report on the state of dataset content.

subdatasets(*[, dataset, state, fulfilled, ...])

Report subdatasets and their properties.

uninstall(*[, dataset, recursive, check, ...])

DEPRECATED: use the drop command

unlock(*[, dataset, recursive, recursion_limit])

Unlock file(s) of a dataset

update(*[, sibling, merge, how, how_subds, ...])

Update a dataset from a sibling.

wtf(*[, sensitive, sections, flavor, decor, ...])

Generate a report about the DataLad installation and configuration



Get a ConfigManager instance for a dataset's configuration


Identifier of the dataset.


path to the dataset


pathobj for the dataset


Get an instance of the version control system/repo for this dataset, or None if there is none yet (or none anymore).