Description¶

Dataverse command-line interface. It uses the Dataverse API and in some cases direct access to the Dataverse database. The commands are implemented using dans-dataverse-client-lib.

dv dataset-edit-metadata¶

For bulk changes on the metadata of multiple datasets in a Dataverse instance, you can use dv dataset-edit-metadata. This command uses the Edit metadata endpoint, which allows you to add or replace metadata fields in a dataset. For typical usage, the command takes a CSV file as input.

Columns¶

The columns fall into several categories:

• Dataset identifier
  • datasetId - the ID or PID of the dataset, a PID must contain the protocol, e.g., doi:10.5072/FK2/123456
• Preconditions
  • expectedState - the expected state of the dataset version to be edited, one of draft, released, deacessioned or a negation of one of those, e.g., not released
  • expectInReview - whether the dataset is expected to be in review: yes or no
• Processing instructions
  • publishVersion - whether to publish the dataset after editing, one of minor, major or leave-draft.
  • replace - whether to replace existing values or add to them: yes or no
• Field values

Field types¶

The names of the field value columns depend on the metadata blocks that are available.

• A primitive field or controlled vocabulary field is specified simply by the field identifier, e.g., myField.
• If multiple values are to be added, the field identifier is indexed, e.g., myField[1]. myField[2], etc.
• A compound field is specified by the field identifier of the parent field, followed by a dot and the name of the child field, e.g., myField.myChildField. Note that in many cases the name of the child field refers to the parent field by convention, e.g., authorName, but you still have to prefix it with the parent field identifier, e.g., author.authorName.
• If multiple values of a compound field are to be added, the parent field identifier is indexed, e.g., myField[1].myChildField, myField[2].myChildField, etc.

Running without a CSV input file¶

When testing, it can be useful to specify a row directly on the command-line, without going to the trouble of first saving it into a CSV file. The columns for dataset identifier, preconditions and processing instructions all have their command line equivalent, e.g., --datasetId, --expectedState, etc. Field values can be specified on the command line as trailing arguments in the format <column name>=<value>. Note that you will need to enclose the whole argument in quotes if the value contains spaces, or if the name has an index, e.g., 'myField[1]=longer value with spaces'.

Report¶

When running with a CSV input file, a report will be written to a file in the directory configured in editMetadata.reportsDir in .dv.yml. The --report-dir option can be used to override this location. The report contains a copy of each line with the extra columns result and message. When running without a CSV input file, the report is written to the standard output.