Frequently Asked Questions

• Access to the data
• Account management
• Administration of the Dataverse collections
• Choosing a data repository
• Curation
• Data papers
• Dataset templates
• Datasets
• Dataverse collections
• DOI
• Files
• General terms and conditions
• Harvesting
• Institutional repositories
• Metadata
• Sandbox and production environments
• The generic space
• User access and rights management

Access to the data

• If a reviewer is sent a private URL, will s/he have access to restricted data?

  Yes, a private URL gives access to files in a draft dataset that has not yet been published even if these files are restricted.

• If the data is already stored elsewhere and accompanied by metadata outside of Recherche Data Gouv, can the "Link to Data" metadata point to the page describing the data rather than to the data itself?

  Yes, if the metadata provides additional information to that entered in the Recherche Data Gouv repository.

  Note: If the other repository assigns a DOI, do not create a dataset in the Recherche Data Gouv repository.

• Can we limit access or place an embargo on a dataset or a collection?

  You can restrict access to a file in a dataset. The Recherche Data Gouv repository does not automatically manage the duration of this access restriction (embargo); it is up to the depositor of the dataset to remove the restriction when the time comes. 

  To restrict access to a dataset as a whole, it must not have been published. Access to this dataset will then only be possible to authorized persons or via its private URL.

  Access to a published Dataverse collection cannot be restricted.

• Is it necessary to indicate in "Link to data" a direct link to the data or can we indicate the database from which the dataset is extracted?

  The "Link to data" metadata indicates the direct link to the data. A link to a database will not allow to find precisely the data described by the dataset. The database from which the dataset is extracted can be mentioned in the metadata "Data sources".

• Does Google collect Recherche Data Gouv's metadata ?

  Yes, the data are indexed by Google Dataset Search.

• Do datasets remain accessible if there is no longer a contact person, in the case of retirement for example?

  The accessibility of the datasets remains the same. However, you should keep a contact person for the dataset (the administrator of the parent collection for example) and also possibly assign rights to modify the dataset or give access to restricted files.

Account management

• How can I find out which collection(s) I have administrator's rights for?

  After logging in, go to your user profile and click on "My data" at the top right of the screen. In the facets, just select "Dataverses" and the "Administrator" role.

• Who validates the registrations of people who create an external account? Is it automatic?

  There is no validation of registrations for people who create an external account. This external account does not open any rights by default.

• I am leaving my current institution. Will I lose access to my datasets?

  You will no longer be able to log into the institutional account linked to the institution you are leaving. If this is the case, you will need to create a new account. This may either be an institutional account associated with your new institution, an ORCID account or an external account. Then, please contact the Repository-registry resource centre (support-recherchedatagouv@inrae.fr) to request the merger of the two accounts and the conservation of the rights to previously created datasets.

• Can an email alias be used for a user account?

  Yes, it is possible to link an email alias to a user account.

Administration of the Dataverse collections

• Is it possible to curate datasets sent via the API?

  Most of the collection functionality is available through the Dataverse APIs, including assigning roles to set up a curation process. A dataset can be sent for review via the API and will only be published if the publish command is executed. A dataset created by the API can also be modified via the user interface.

• Can a collection administrator create an alert to be notified when a dataset has been created for over a year without being published?

  No. However, it IS possible to identify datasets in a collection that were created on a specific date and remain unpublished.

  For example: write the dateOfDeposit:[* TO 2020] as your query then select "unpublished" datasets via the facets (you need to log in to use this facet).

• Is there a graphic charter which needs to be followed for collections?

  No, the Research Data Gouv platform makes no specific recommendations on graphics although some technical constraints are indicated in the collection settings (Theme + Widget menu).

  Nonetheless you may consider it advisable to check the specific documentation of the institutional space you are working in to see if there are any particular recommendations.

• Can optional metadata be made obligatory in nature? If this is the case, what happens to datasets that have already been published without the optional metadata being filled in?

  Yes, optional metadata can be made obligatory. This has no impact on datasets that have already been published as long as these are not modified. However when a modification is made, the obligatory metadata needs to be filled in. 

Choosing a data repository

• I need to centralise data from a multi-partner programme. Can I deposit the data produced by all the partners be deposited in the Research Data Gouv Repository?

  Yes, all the data from this kind of project can be deposited in the Recherche Data Gouv repository. If one of the partners possesses an institutional space, then the data should be deposited there in the project's dedicated collection. If this is not the case, the data should be deposited in the generic space.

  If the datasets have already been deposited elsewhere, please indicate the link to these datasets in the "Related datasets" metadata.

  Please remember to specify each producer in the "Producer" metadata.

• Can I deposit my data in another repository

  Yes, you are strongly recommended to deposit your research data in a reference repository for your thematic community. If this is not possible, the data can be deposited in a repository housed by your institution or in the Research Data Gouv repository.

• Who can deposit data in the Recherche Data Gouv repository?

  Only research data produced by research teams with at least one collaborator affiliated to an institution belonging to the French public research community are accepted by right in the Recherche Data Gouv repository.

• What criteria does a dataset need to meet to be accepted for publication in Recherche Data Gouv repository?

  A dataset needs to meet the following criteria to be published in the Recherche Data Gouv repository:

  • The data has to have been produced in a research context;

  • The data should preferably be structured and deposited in a machine-readable format.

Curation

• As the depositor, can I still take action on a dataset which is being reviewed?

  No. Depositors can take no action on their datasets while these are being reviewed.

• Is it possible to curate datasets sent via the API?

  Most of the collection functionality is available through the Dataverse APIs, including assigning roles to set up a curation process. A dataset can be sent for review via the API and will only be published if the publish command is executed. A dataset created by the API can also be modified via the user interface.

Data papers

• If a reviewer is sent a private URL, will s/he have access to restricted data?

  Yes, a private URL gives access to files in a draft dataset that has not yet been published even if these files are restricted.

• Which should be published first – the article or the data paper?

  Either may be published first. This is a question of scientific strategy and is the authors' responsibility. If authors are concerned about the use of the data, the data paper can be written first but published after the scientific article has been published. If in doubt, please contact the editor.

• Can a data paper be generated from a collection (and not just from a dataset)?

  No, generating a data paper using the dedicated functionality requires you to use a dataset's DOI.

Dataset templates

• As there is a "Recherche Data Gouv" template, is there any point in creating or using other models?

  The Recherche Data Gouv dataset template only pre-fills in the information linked to the disseminator and the Open/Etalab licence. Collection administrators are strongly recommended to create new and more complete templates to make it easier to fill in as many metadata fields as possible when depositing, for example in the context of a project or thematic collection.

  To create a dataset template, please consult: 

  https://recherche.data.gouv.fr/en/category/27/guide/modifying-the-parameters-of-a-collection#Creating-dataset-templates

• Can a template be applied to an existing dataset?

  No, a template can only be used to create new datasets.

• Is it possible to have several templates for the same collection and for the same dataset?

  Collection: Yes, it is possible to have several templates in a collection including those of the parent collection (e.g. the INRAE template).

  Dataset: No, it is not possible to choose several templates for a dataset or to change your template once the dataset has been created.

• Are there templates for collections or datasets?

  Dataset templates are created by the administrator in a collection and apply to the datasets created in that collection.

Datasets

• Can I move a dataset without associating it with a collection?

  No, a dataset is always created in a collection. This collection can be the generic space or a collection in one of the institutional spaces.

• What is the recommended timeframe for the publication of a dataset?

  Depositors are recommended to publish their datasets as soon as possible because the Research Data Gouv repository aims to serve as a repository for data access.

  It is advisable not to take more than one year to publish data which are not linked to a publication. After this period, an administrator of the collection the data is linked to may contact the depositor to alert him or her.

• Can I delete a dataset that has not been published?

  Yes, a dataset can be deleted as long as it has not yet been published. Only draft versions can be deleted.

• What is the maximum size for a dataset?

  There is no limit to the size of a dataset.

  Please see "What is the size of an institutional space?" and "What is the maximum size for a file?".

• Is my dataset still accessible if the files are deleted?

  The accessibility of a dataset does not depend on files being present. However, if all the files in a dataset are deleted then you should specify the link to the data in the "Link to data" metadata.

Dataverse collections

• Can a collection be linked to more than one parent collection?

  A collection can only be linked to one parent collection, the collection in which it was created. One collection can be linked to another so that it appears as a sub-collection. To link a collection, please get in touch with the Repository-registry resource centre: support-recherchedatagouv@inrae.fr.

• Can a dataset be attached to a collection a posteriori?

  Yes, but only the Recherche Data Gouv repository resource center can move a dataset from one collection to another. It is better to make a link than a move.

• Can collections be moved?

  Only the functional administrators of the Recherche Data Gouv repository can do this and not the administrator of a given collection. This means you should think carefully about the positioning of a collection before creating it.

  It is possible to link a collection to another collection to make it appear as a sub-collection but the contents of the latter will not be visible in the target collection.

  To move or link a collection please contact the Repository-registry resource centre at support-recherchedatagouv@inrae.fr.

• What is the maximum size for a collection?

  There is no maximum size for a collection other than the volume of the institutional space that contains it. Please see "What is the size of an institutional space?"

DOI

• Is a DOI standardized? unique?

  Yes, the DOI is based on the Handle identifier system and is ISO certified (ISO 26324, Digital Object Identifier System). Uniqueness is ensured by the fact that suffixes are unique for a given prefix.

• Should I indicate the DOI of a Collective Scientific Infrastructure in a dataset?

  When a Scientific Collective Infrastructure (SCI) has contributed to the dataset, it can be indicated in the Contributor metadata by specifying the "Type" and choosing "DOI" in "Contributor Identifier Scheme". The DOI itself is entered in "Contributor Identifier".

• When publishing a dataset via an API on the sandbox environment, is it also saved on DataCite?

  The sandbox publishes datasets only on the DataCite test environment, which is a closed system: it is the only way to directly find the resource corresponding to the DOI. The DOI does not lead to any landing page for the dataset.

• Is the DOI maintained if the dataset is not published quickly?

  The DOI of a dataset is generated and reserved when the dataset is created. It is activated when the dataset is published.

  As a reminder, data deposited in the Research Data Gouv warehouse must be published within 12 months.

• Can I create a dataset in the Recherche Data Gouv repository if it has already been deposited in another repository that also assigns DOIs?

  No, it is not recommended to create a new dataset in the Recherche Data Gouv repository as this would mean a new DOI would be assigned to it. Please see the depositing guide:

  https://recherche.data.gouv.fr/en/category/9/guide/before-depositing

• If I update my dataset is the DOI modified?

  No, the DOI stays the same if the version changes. Please see the depositing guide:

  https://recherche.data.gouv.fr/en/category/9/guide/modifying-and-managing-versions-of-a-published-dataset

Files

• Do I need to re-publish every time a file is modified?

  A file modification creates a new draft dataset. A new publication of this dataset makes such changes public and updates the version.

• What should I do if I realise my file has not been ingested correctly?

  Please consult the guide to identify the cause of the error (https://recherche.data.gouv.fr/en/category/9/guide/depositing-a-dataset#Tabulated+data+files), then replace the problem file with the corrected file.

• How to upload a large amount of files at once?

  The repository interface of the Recherche Data Gouv repository limits the upload to 1000 files at a time. Beyond that, it is possible to use the DVUploader tool or the native API.

• Which formats can files be exported in?

  Files can be exported in their original format. Tabular data can also be exported in Rdata and tabulated formats.

• How do we anonymize personal data?

  It is possible to use anonymization tools, such as Amnesia (https://amnesia.openaire.eu/), which allows anonymized data to be send directly to the Recherche Data Gouv repository.

• Is there a naming convention for files, datasets or Dataverse collections?

  Recommendations for naming and organizing data files are available on the Doranum website ((https://doranum.fr/tags/nommage-fichier/). In addition, specific recommendations can be proposed by the administrator of a Dataverse collection and integrated into the data management plan associated with it.

• Can a file filed in the Recherche Data Gouv repository be associated with multiple datasets?

  A file can only be deposited in one dataset. However, it is possible to refer to a file deposited in another dataset via the "Related Datasets" metadata.

• Can we point directly to a file (and not to its landing page)?

  This is not possible from the user interface. You have to use the native API - Accessing (downloading) files - by indicating the DOI of the file with the command https://entrepot.recherche.data.gouv.fr/api/datafile/:persistentId/?persistentId=doi:{DOI du fichier}

• Can I update a file without having to delete it and making a new deposit?

  It is not possible to modify the contents of a file in the user interface. To update a file, it needs to be replaced. Please note that this would involve a version change. The dataset must be republished for the new file to be accessed.

• Can I upload zipped files or archives?

  Yes, all file formats are accepted. If your dataset has a large tree structure, we strongly recommend you use DVUploader : https://recherche.data.gouv.fr/en/category/33/guide/dv-uploader-1 to upload all your files and the tree structure rather than an archive.

• What is the maximum size for a file?

  The maximum size for file is 50 Go.

• Is it possible to sort files in a dataset by number of downloads?

  No, it is not possible to sort files by number of downloads. You can sort the files in a dataset by name, date of deposit, size and category (file type).

• Can Excel files be ingested with multiple sheets?

  No, if the file contains several sheets, it can be uploaded to the Recherche Data Gouv repository but only the first sheet is ingested and thus transformed into a .tab file. 

  In order to achieve good ingestion, it is therefore advisable to upload only files containing a single tab with the variables on the first line (column headers) and one observation per line (see Case of tabulated data files).

General terms and conditions

• How long will the data be stored?

  The secure hosting and availability of the data is guaranteed for a minimum of 5 years renewable after publication. This period is a minimum, not a maximum period of data retention, which may vary depending on the dataset.

• I need to centralise data from a multi-partner programme. Can I deposit the data produced by all the partners be deposited in the Research Data Gouv Repository?

  Yes, all the data from this kind of project can be deposited in the Recherche Data Gouv repository. If one of the partners possesses an institutional space, then the data should be deposited there in the project's dedicated collection. If this is not the case, the data should be deposited in the generic space.

  If the datasets have already been deposited elsewhere, please indicate the link to these datasets in the "Related datasets" metadata.

  Please remember to specify each producer in the "Producer" metadata.

• How long is my dataset accessible?

  Secure hosting and availability of the data is guaranteed by the Recherche Data Gouv repository representative for a minimum of 5 years renewable after the dataset is published.

  The link between DOI and dataset description page in the Recherche Data Gouv repository is guaranteed. This follows the rules set by DataCite, the agency managing the DOI. 

Harvesting

• Is it possible to harvest Zenodo?

  No, Zenodo harvesting is not supported by the current version of the Dataverse software.

Institutional repositories

• How to request the creation of an institutional repository?

  Information on how to request the creation of an institutional repository is available on the "Joining the ecosystem" page.

• What is the size of an institutional space?

  An institutional space has 5 TB when it is created.

• What is the volume of an institutional space?

  By default, the volume of an institutional space is 5To.

Metadata

• Where does the metadata offered by the Recherche Data Gouv repository come from?

  Most of the metadata comes from Dataverse software (see Metadata References) and is compliant with the Documentation Initiative (DDI), Dublin Core, DataCite and ISA-Tab standards. Other metadata has been created for the specific needs of an institution, such as the Semantic resource block developed by INRAE.

  The values of some metadata can come from external repositories or from the Recherche Data Gouv repository.

• Is a DOI standardized? unique?

  Yes, the DOI is based on the Handle identifier system and is ISO certified (ISO 26324, Digital Object Identifier System). Uniqueness is ensured by the fact that suffixes are unique for a given prefix.

• Should I indicate the DOI of a Collective Scientific Infrastructure in a dataset?

  When a Scientific Collective Infrastructure (SCI) has contributed to the dataset, it can be indicated in the Contributor metadata by specifying the "Type" and choosing "DOI" in "Contributor Identifier Scheme". The DOI itself is entered in "Contributor Identifier".

• What is the difference between an author and a contributor?

  The author is the person responsible for the dataset and is cited in the dataset. The authorName metadata associated with the DOI is obligatory in the Recherche Data Gouv repository.

  A contributor is a person or organisation who/which took part in the collection, management, distribution of the dataset or contributed to it in any other way. Contributors do not appear in dataset citations.

• What should be filled into the "Journal Metadata" block?

  Information about the journal in which the data is published should be entered here.

• What is DDI?

  The Data Documentation Initiative (DDI) is a standard created by the DDI Alliance which enables users to document survey and observational data in the social, behavioural, economic and health sciences.

• Can the metadata be completed after a dataset is published?

  Yes, it is perfectly possible to complete the metadata for a dataset after this has been published. A new (minor or major) version will thus be created and must be published in turn.

• Is English compulsory for all metadata, for example "Description"?

  No, the choice of language depends on the target audience for your publication. Different "Description" metadata can be created by specifying the language of each.

• What does "Incomplete metadata" mean?

  The label "incomplete metadata" appears when, as it stands, one or more mandatory metadata fields are missing. This is due to a change in parameters, and prevents the dataset from being published or upgraded until the field(s) have been completed.

• Can we add a list of authors or contributors via a file?

  No, this functionality does not exist from the user interface. It is however possible to use APIs to perform this action. For more information, contact the Recherche Data Gouv repository resource center.

• Can we integrate our own metadata, for example in the case of a project that has its own referential?

  It is not possible for a user to add metadata to the form. It is possible, however, to upload an documentation file with the data file containing the specific metadata.

  To suggest that metadata be added to the input form, contact the Recherche Data Gouv repository resource center..

• If the data is already stored elsewhere and accompanied by metadata outside of Recherche Data Gouv, can the "Link to Data" metadata point to the page describing the data rather than to the data itself?

  Yes, if the metadata provides additional information to that entered in the Recherche Data Gouv repository.

  Note: If the other repository assigns a DOI, do not create a dataset in the Recherche Data Gouv repository.

• Which identifier is best suited in the Recherche Data Gouv repository to identify a publisher?

  The ORCID identifier is recommended in the second French Plan for Open Science to identify authors and contributors.

  However, you can use other identifiers in the Recherche Data Gouv repository, knowing that it is only possible to indicate one identifier per author or contributor.

Sandbox and production environments

• Can we import a dataset created in the sandbox into the production environment?

  No, these two environments are independent and there is no possibility to export a dataset from one to import it into the other.

• Is there a testing area to practice on before publishing a dataset or collection in the Recherche Data Gouv repository?

  Yes there is a sandbox https://demo.recherche.data.gouv.fr/.

  All authorised users can create datasets or collections while in the "Travaux Pratiques"collection.

• What permissions do I have in the sandbox environment?

  When you create an account in the sandbox environment, you then have the default permissions of a dataset and collection creator in the “Travaux Pratiques” collection.

• When publishing a dataset via an API on the sandbox environment, is it also saved on DataCite?

  The sandbox publishes datasets only on the DataCite test environment, which is a closed system: it is the only way to directly find the resource corresponding to the DOI. The DOI does not lead to any landing page for the dataset.

The generic space

• Can a user affiliated to an institution with its own institutional space create a dataset in the Espace générique?

  No, a dataset can only be deposited in the generic space if neither the depositor or one of her/his team members has no access to an institutional space.

• Can I deposit my data in the Espace générique while waiting for an institutional space to be set up or is it better to wait?

  Yes, it is possible to deposit data in the generic space rather than waiting for an institutional space to be set up. However, if one of the team members has an institutional space then the dataset must be deposited there.

User access and rights management

• Is it necessary to set new permissions whenever a new collection is created ?

  Not necessarily, by default the creator of a collection is the only user having rights within it. These rights could be modified via the menu "Edit > Permissions".

• Can a role be assigned to all members of a project at the same time?

  Yes, it is possible to create a group made up of all the members of the project and then assign the same role to the group created in this way. (https://recherche.data.gouv.fr/en/category/27/guide/modifying-the-parameters-of-a-collection#Groups).