Context
By default, File Explorer and the S3 section of service or process launch forms use the storage configuration selected in Project Settings → S3 Configurations.
You can change this configuration or connect an external S3-compatible bucket. If several buckets are configured, the one with Use for EDITO explorers enabled is displayed in File Explorer.
📌 Note: S3 object storage is not a traditional filesystem. The folders shown in File Explorer are prefixes in object keys, not independent directories (see Understand EDITO storage).
This article focuses on data that is already accessible from EDITO. To transfer data from your own computer, workstation or HPC cluster, see Transfer data between your computer or cluster and EDITO.
Manage objects with File Explorer
Open File Explorer in the Datalab to manage the selected bucket graphically. File Explorer lets you:
Browse your bucket as in regular filesystem
Upload and download files
(1)and(2)Rename files
Create a new directory
(3)Delete files
(4)Share files
(5)and(6)Open files with cloud-native formats in the Datalab’s Data Explorer
(7)Check the CLI of the File Explorer
(8)Temporarily or permanently share files or folders publicly
(9)
⚠️ Some features of the File Explorer Graphical User Interface (GUI) are experiencing issues: we recommend that you do everything through Command Line Interface (CLI).
For public-access settings, see Share files and datasets from EDITO storage.
Quickly open files in Jupyter
Select one or more files or prefixes in File Explorer, then choose Open in Jupyter Notebook.
EDITO launches the Jupyter-python-ocean-science service and copies the selected content into the service's working directory.
⚠️ The copy inside Jupyter is separate from the object stored in S3. Editing the local copy does not update the original object. Before stopping the service, copy any modified or newly created files back to permanent storage.
Use MinIO Client inside a running service
In an EDITO service such as Jupyter, the S3 endpoint and credentials are normally injected automatically. The MinIO Client alias s3 is also configured for you.
Open a terminal and inspect the available variables if needed:
env | grep -E 'AWS|S3|MC_HOST'
You should not need to run mc alias set when working inside a correctly configured EDITO service.
Identify your bucket
A personal bucket is normally named oidc-[YOUR_USERNAME].
A project or external bucket may use another name or a configured working-directory path. Check Project Settings → S3 Configurations if you are unsure.
In the following examples, replace:
<WORKING_DIRECTORY_PATH>with the bucket name;path/to/objectwith the object key;/local/pathwith a path inside the running service.
List bucket content
mc ls s3/<WORKING_DIRECTORY_PATH>
List a specific prefix:
mc ls s3/<WORKING_DIRECTORY_PATH>/path/to/prefix/
For personal storage, for example:
mc ls s3/oidc-[YOUR_USERNAME]
Copy an object from S3 to the running service
mc cp s3/<WORKING_DIRECTORY_PATH>/path/to/object /local/path/to/object
Example for personal storage:
mc cp s3/oidc-[YOUR_USERNAME]/path/to/my_file /local/path/to/my_file
Copy all objects under a prefix to the running service
mc cp --recursive s3/<WORKING_DIRECTORY_PATH>/path/to/prefix/ /local/path/to/folder/
Example for your personal storage:
mc cp --recursive s3/oidc-[YOUR_USERNAME]/path/to/my_folder/ /local/path/to/my_folder/
Copy an object from the service back to S3
mc cp /local/path/to/object s3/<WORKING_DIRECTORY_PATH>/path/to/object
Example for your personal storage:
mc cp /local/path/to/my_file s3/oidc-[YOUR_USERNAME]/path/to/my_file
Copy a local directory back to S3
mc cp --recursive /local/path/to/folder/ s3/<WORKING_DIRECTORY_PATH>/path/to/prefix/
Use this before stopping a Jupyter or another service if the local files must persist.
Organize your storage with mc
Create a folder-like path
S3 object storage does not use folders in the same way as a traditional file system. What appears as a folder in File Explorer or MinIO Client is actually a prefix in the object key.
You therefore do not need to create a folder before uploading a file. Simply include the desired path in the destination object name:
mc cp local_file.nc s3/<WORKING_DIRECTORY_PATH>/new_folder/local_file.nc
After the upload, new_folder/ will appear as a folder-like path in the storage browser.
You can use the same approach to create several levels:
mc cp local_file.nc s3/<WORKING_DIRECTORY_PATH>/results/2026/local_file.nc
This will make the object appear under:
results/
└── 2026/
└── local_file.nc
📌 Note: mc mb creates a new bucket (folder-like), not a folder inside an existing bucket. Use it only when you actually need to create a bucket and have permission to do so:
mc mb s3/<WORKING_DIRECTORY_PATH>/new_bucket/
For most EDITO users, creating a new bucket is not necessary: use your existing personal or project bucket and organize its contents with prefixes.
Copy an object inside a bucket
mc cp s3/<WORKING_DIRECTORY_PATH>/source/path/my_file.nc s3/<WORKING_DIRECTORY_PATH>/destination/path/my_file.nc
Copy all objects under a prefix
mc cp --recursive s3/<WORKING_DIRECTORY_PATH>/source_prefix/ s3/[BUCKET]/destination_prefix/
The source and destination can also use different configured aliases or buckets.
Move or rename an object
mc mv s3/<WORKING_DIRECTORY_PATH>/old_path/my_file.nc s3/<WORKING_DIRECTORY_PATH>/new_path/my_file.nc
Move or rename a prefix
mc mv --recursive s3/<WORKING_DIRECTORY_PATH>/old_prefix/ s3/<WORKING_DIRECTORY_PATH>/new_prefix/
For large prefixes, remember that object storage may implement this operation as a copy followed by deletion.
Check your storage
mc du s3/<WORKING_DIRECTORY_PATH>
Delete one object
mc rm s3/<WORKING_DIRECTORY_PATH>/path/to/my_file.nc
Delete all objects under a prefix
Preview the scope first:
mc rm --recursive --force --dry-run s3/<WORKING_DIRECTORY_PATH>/path/to/prefix/
Then run the deletion:
mc rm --recursive --force s3/<WORKING_DIRECTORY_PATH>/path/to/prefix/
⚠️ mc rm --recursive --force is irreversible unless the bucket has an independent versioning or backup policy. Check the alias, bucket and prefix carefully before running it.
Use File Explorer CLI snippets
File Explorer can display command-line snippets for some graphical actions. These snippets are useful for confirming the current alias, bucket and path before reproducing an operation in a terminal.
For additional commands and options, refer to the MinIO Client documentation linked from File Explorer.
Good practices
Keep object keys descriptive and stable.
Use clear versioned prefixes for published or reproducible datasets.
Do not begin object keys with
/.Check the available quota before copying a large dataset.
mchandles multipart upload automatically for large files; no extra configuration is normally required.Test destructive commands on a small prefix before applying them to a complete dataset.
Copy important service outputs to S3 before stopping or deleting the service.
For an officially published dataset, avoid replacing objects in place. Create a new version and update the associated metadata instead.
What's next?
If you have any questions, problems, or suggestions, please feel free to contact us via chat using the widget available at the bottom right of the page.