Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Editing datasets","local":"editing-datasets","sections":[{"title":"Edit using the Hub UI","local":"edit-using-the-hub-ui","sections":[],"depth":2},{"title":"Using the huggingface_hub client library","local":"using-the-huggingfacehub-client-library","sections":[],"depth":2},{"title":"Integrated libraries","local":"integrated-libraries","sections":[],"depth":2},{"title":"Only upload the new data","local":"only-upload-the-new-data","sections":[],"depth":2},{"title":"Optimized Parquet editing","local":"optimized-parquet-editing","sections":[],"depth":2},{"title":"Streaming","local":"streaming","sections":[],"depth":2}],"depth":1}"> | |
| <link href="/docs/hub/pr_2437/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/entry/start.5e3ed1fd.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/scheduler.258d2a4d.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/singletons.2d0c91e1.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/index.c8b82093.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/paths.d360121b.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/entry/app.93e7704b.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/preload-helper.a4507a26.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/index.421344fd.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/nodes/0.bedbef6e.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/each.e59479a4.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/nodes/40.3d88e7b2.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/CopyLLMTxtMenu.9c2a67a1.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.9bb92958.js"> | |
| <link rel="modulepreload" href="/docs/hub/pr_2437/en/_app/immutable/chunks/CodeBlock.619ec4e3.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Editing datasets","local":"editing-datasets","sections":[{"title":"Edit using the Hub UI","local":"edit-using-the-hub-ui","sections":[],"depth":2},{"title":"Using the huggingface_hub client library","local":"using-the-huggingfacehub-client-library","sections":[],"depth":2},{"title":"Integrated libraries","local":"integrated-libraries","sections":[],"depth":2},{"title":"Only upload the new data","local":"only-upload-the-new-data","sections":[],"depth":2},{"title":"Optimized Parquet editing","local":"optimized-parquet-editing","sections":[],"depth":2},{"title":"Streaming","local":"streaming","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"><div class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg></span> <span>Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg></button></div> </div> <h1 class="relative group"><a id="editing-datasets" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#editing-datasets"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Editing datasets</span></h1> <p data-svelte-h="svelte-14evjus">The <a href="https://huggingface.co/datasets" rel="nofollow">Hub</a> enables collaborative curation of community and research datasets. We encourage you to explore the datasets available on the Hub and contribute to their improvement to help grow the ML community and accelerate progress for everyone. All contributions are welcome!</p> <p data-svelte-h="svelte-1fojn2">Start by <a href="https://huggingface.co/join" rel="nofollow">creating a Hugging Face Hub account</a> if you don’t have one yet.</p> <h2 class="relative group"><a id="edit-using-the-hub-ui" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#edit-using-the-hub-ui"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Edit using the Hub UI</span></h2> <blockquote class="warning" data-svelte-h="svelte-kt6wx2"><p>This feature is only available for CSV, TSV, and Parquet datasets for now.</p></blockquote> <p data-svelte-h="svelte-1ur7487">The Hub’s web interface allows users without any technical expertise to edit a dataset.</p> <p data-svelte-h="svelte-fuyrts">Open the dataset page and navigate to the <strong>Data Studio</strong> tab to begin editing.</p> <div class="flex justify-center" data-svelte-h="svelte-s120s5"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/data_studio_button-min.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/data_studio_button_dark-min.png"></div> <p data-svelte-h="svelte-bb2vkp">Click on <strong>Toggle edit mode</strong> to enable dataset editing.</p> <div class="flex justify-center" data-svelte-h="svelte-lwt9p1"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/toggle_edit_button-min.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/toggle_edit_button_dark-min.png"></div> <div class="flex justify-center" data-svelte-h="svelte-1o02ujd"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/edit_cell_button-min.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/edit_cell_button_dark-min.png"></div> <p data-svelte-h="svelte-6i50fi">Edit as many cells as you want and finally click <strong>Commit</strong> to commit your changes and leave a commit message.</p> <div class="flex justify-center" data-svelte-h="svelte-u5mgd3"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/commit_button-min.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/commit_button_dark-min.png"></div> <div class="flex justify-center" data-svelte-h="svelte-7fk0qf"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/commit_message-min.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-edit/commit_message_dark-min.png"></div> <h2 class="relative group"><a id="using-the-huggingfacehub-client-library" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#using-the-huggingfacehub-client-library"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Using the huggingface_hub client library</span></h2> <p data-svelte-h="svelte-yzf7y1">The <code>huggingface_hub</code> library can manage Hub repositories including editing datasets.</p> <p data-svelte-h="svelte-pesmie">For example here is how to edit a CSV file using the <a href="https://huggingface.co/docs/huggingface_hub/en/guides/hf_file_system" rel="nofollow">Hugging Face FileSystem API</a>:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> hffs | |
| path = <span class="hljs-string">f"datasets/<span class="hljs-subst">{repo_id}</span>/data.csv"</span> | |
| <span class="hljs-keyword">with</span> hffs.<span class="hljs-built_in">open</span>(path, <span class="hljs-string">"r"</span>) <span class="hljs-keyword">as</span> f: | |
| content = f.read() | |
| edited_content = content.replace(<span class="hljs-string">"foo"</span>, <span class="hljs-string">"bar"</span>) | |
| <span class="hljs-keyword">with</span> hffs.<span class="hljs-built_in">open</span>(path, <span class="hljs-string">"w"</span>) <span class="hljs-keyword">as</span> f: | |
| f.write(edited_content)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1j4uxrl">You can also apply edit locally on your disk and commit the changes:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> hf_hub_download, upload_file | |
| local_path = hf_hub_download(repo_id=repo_id, path_in_repo= <span class="hljs-string">"data.csv"</span>, repo_type=<span class="hljs-string">"dataset"</span>) | |
| <span class="hljs-keyword">with</span> <span class="hljs-built_in">open</span>(path, <span class="hljs-string">"r"</span>) <span class="hljs-keyword">as</span> f: | |
| content = f.read() | |
| edited_content = content.replace(<span class="hljs-string">"foo"</span>, <span class="hljs-string">"bar"</span>) | |
| <span class="hljs-keyword">with</span> <span class="hljs-built_in">open</span>(path, <span class="hljs-string">"w"</span>) <span class="hljs-keyword">as</span> f: | |
| f.write(edited_content) | |
| upload_file(repo_id=repo_id, path_in_repo=local_path, repo_type=<span class="hljs-string">"dataset"</span>)<!-- HTML_TAG_END --></pre></div> <blockquote class="tip" data-svelte-h="svelte-19cwwnp"><p>To have the entire dataset repository locally and edit many files at once, use <code>snapshot_download</code> and <code>upload_folder</code> instead of <code>hf_hub_download</code> and <code>upload_file</code></p></blockquote> <p data-svelte-h="svelte-1pwsgxh">Visit <a href="/docs/huggingface_hub/index">the client library’s documentation</a> to learn more.</p> <h2 class="relative group"><a id="integrated-libraries" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#integrated-libraries"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Integrated libraries</span></h2> <p data-svelte-h="svelte-1mys076">If a dataset on the Hub is compatible with a <a href="./datasets-libraries">supported library</a>, loading, editing, and pushing the dataset takes just a few lines.</p> <p data-svelte-h="svelte-wmqi3n">Here is how to edit a CSV file with Pandas:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> pandas <span class="hljs-keyword">as</span> pd | |
| <span class="hljs-comment"># Load the dataset</span> | |
| df = pd.read_csv(<span class="hljs-string">f"hf://datasets/<span class="hljs-subst">{repo_id}</span>/data.csv"</span>) | |
| <span class="hljs-comment"># Edit</span> | |
| df = df.apply(...) | |
| <span class="hljs-comment"># Commit the changes</span> | |
| df.to_csv(<span class="hljs-string">f"hf://datasets/<span class="hljs-subst">{repo_id}</span>/data.csv"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-o56se8">Libraries like Polars and DuckDB also implement the <code>hf://</code> protocol to read, edit and write files on Hugging Face. And other libraries are useful to edit datasets made of many files like Spark, Dask or 🤗 Datasets. See the full list of supported libraries <a href="./datasets-libraries">here</a></p> <p data-svelte-h="svelte-2uoc93">For information on accessing the dataset on the website, you can click on the “Use this dataset” button on the dataset page to see how to do so. | |
| For example, <a href="https://huggingface.co/datasets/knkarthick/samsum?library=datasets" rel="nofollow"><code>samsum</code></a> shows how to do so with 🤗 Datasets below.</p> <div class="flex justify-center" data-svelte-h="svelte-1ach5zb"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-usage.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-usage-dark.png"></div> <div class="flex justify-center" data-svelte-h="svelte-1rm4drj"><img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-usage-modal.png"> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/datasets-usage-modal-dark.png"></div> <h2 class="relative group"><a id="only-upload-the-new-data" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#only-upload-the-new-data"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Only upload the new data</span></h2> <p data-svelte-h="svelte-1u44vdp">Hugging Face’s storage is powered by <a href="https://huggingface.co/docs/hub/en/xet" rel="nofollow">Xet</a>, which uses chunk deduplication to make uploads more efficient. | |
| Unlike traditional cloud storage, Xet doesn’t require the entire dataset to be re-uploaded to commit changes. | |
| Instead, it automatically detects which parts of the dataset have changed and instructs the client library only to upload the updated parts. | |
| To do that, Xet uses a smart algorithm to find chunks of 64kB that already exist on Hugging Face.</p> <p data-svelte-h="svelte-e8xbr0">Let’s see our previous example with Pandas:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> pandas <span class="hljs-keyword">as</span> pd | |
| <span class="hljs-comment"># Load the dataset</span> | |
| df = pd.read_csv(<span class="hljs-string">f"hf://datasets/<span class="hljs-subst">{repo_id}</span>/data.csv"</span>) | |
| <span class="hljs-comment"># Edit part of the dataset</span> | |
| df = df.apply(...) | |
| <span class="hljs-comment"># Commit the changes</span> | |
| df.to_csv(<span class="hljs-string">f"hf://datasets/<span class="hljs-subst">{repo_id}</span>/data.csv"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1my2lh2">This code first loads a dataset and then edits it. | |
| Once the edits are done, <code>to_csv()</code> materializes the file in memory, chunks it, asks Xet which chunks are already on Hugging Face and which chunks have changed, and then uploads only the new data.</p> <h2 class="relative group"><a id="optimized-parquet-editing" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#optimized-parquet-editing"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Optimized Parquet editing</span></h2> <p data-svelte-h="svelte-joxyj">The amount of data to upload depends on the edits and the file structure.</p> <p data-svelte-h="svelte-178yn32">The Parquet format is columnar and compressed at the page level (pages are around ~1MB). | |
| We optimized Parquet for Xet with <a href="https://huggingface.co/blog/parquet-cdc" rel="nofollow">Parquet Content Defined Chunking</a>, which ensures unchanged data generally result in unchanged pages.</p> <p data-svelte-h="svelte-1wvjvek">For example, this code uploads the content of <code>df</code> and then for <code>edited_df</code> the upload is faster since it only uploads the chunks that changed:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> pandas <span class="hljs-keyword">as</span> pd | |
| df.to_parquet( | |
| <span class="hljs-string">"hf://datasets/username/my_dataset/imdb.parquet"</span>, | |
| <span class="hljs-comment"># Optimize for Xet</span> | |
| use_content_defined_chunking=<span class="hljs-literal">True</span>, | |
| write_page_index=<span class="hljs-literal">True</span>, | |
| ) | |
| edited_df = ... <span class="hljs-comment"># e.g. with added/modified/removed rows or columns</span> | |
| edited_df.to_parquet( | |
| <span class="hljs-string">"hf://datasets/username/my_dataset/imdb.parquet"</span>, | |
| <span class="hljs-comment"># Optimize for Xet</span> | |
| use_content_defined_chunking=<span class="hljs-literal">True</span>, | |
| write_page_index=<span class="hljs-literal">True</span>, | |
| )<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1psrwcu">Chunks are ~64kB and Parquet saves data column per column, so in practice this is what happens when editing an Optimized Parquet file:</p> <ul data-svelte-h="svelte-1scito"><li>add a new column -> only the chunks of the new column are uploaded</li> <li>add/edit/delete a row -> one chunk per column is uploaded</li></ul> <p data-svelte-h="svelte-nyjwuh">And in addition to this, the chunks of the Parquet footer containing metadata are also uploaded.</p> <p data-svelte-h="svelte-kb0d0x">Check out if your library supports Optimized Parquet in the <a href="./datasets-libraries">supported libraries</a> page.</p> <h2 class="relative group"><a id="streaming" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#streaming"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Streaming</span></h2> <p data-svelte-h="svelte-7o2tfc">For big datasets, libraries with dataset streaming features for end-to-end streaming pipelines are recommended. | |
| In this case, the dataset processing runs progressively as the old data arrives and the new data is uploaded to the Hub.</p> <p data-svelte-h="svelte-5cay8u">Check out if your library supports streaming in the <a href="./datasets-libraries">supported libraries</a> page.</p> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/hub-docs/blob/main/docs/hub/datasets-editing.md" target="_blank"><svg class="mr-1" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg> <span data-svelte-h="svelte-zjs2n5"><span class="underline">Update</span> on GitHub</span></a> <p></p> | |
| <script> | |
| { | |
| __sveltekit_ummgov = { | |
| assets: "/docs/hub/pr_2437/en", | |
| base: "/docs/hub/pr_2437/en", | |
| env: {} | |
| }; | |
| const element = document.currentScript.parentElement; | |
| const data = [null,null]; | |
| Promise.all([ | |
| import("/docs/hub/pr_2437/en/_app/immutable/entry/start.5e3ed1fd.js"), | |
| import("/docs/hub/pr_2437/en/_app/immutable/entry/app.93e7704b.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 40], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 32.7 kB
- Xet hash:
- c0ee01c2641cef70ea7e33d741214e70e642bc34bb9deb9529a5fd72e5d5eb2d
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.