Column names are incorrect

Issue description

Import fails with error that a column name is not valid for this artifact type, or that a mandatory column is missing, but in fact all the columns in the file are named correctly and none of the mandatory columns is missing.

Workaround

The headers might contain some whitespace characters that are not visible in the spreadsheet or even in the text editor. The whitespace character prevents the column name from being recognized.

• Edit the file with with an editor that shows whitespace characters.
• Try deleting the headers and type their names manually from scratch.
• Check whether the file encoding is UTF-8, and not, for example, UTF-8 BOM.

Relationships to nonexistent artifacts are skipped

Issue description

When you define any kind of relationship between two artifacts, the import process succeeds, but relationships to nonexistent artifacts are skipped. Skipped relationships result in errors similar to this:

GIM00015E: Artifact Average balance in column Related Terms is not found in the specified hierarchy.

Workaround

If you plan to import multiple types of artifacts that have relationships between each other, first import all the artifacts, publish the imported artifacts, and then import the same CSV files again, with the Replace all values merge method. The second import processes adds the relationships to the existing artifacts.

Category doesn't exist

Issue description

If a parent category for an artifact is not defined in the CVS file, the artifacts are imported to the [uncategorized] category by default. If a parent category for an artifact is defined in the CSV file, but it doesn't exist in the environment, the import fails. The import process always creates only one type of artifact per import. When importing artifacts other than categories, the parent category can't be created.

Workaround

Before you import artifacts from a CSV file, make sure that the categories which you want to import them to exist in the environment. Create them before you start the import process.

Can't find the artifacts after CSV import

Issue description

The artifacts that are imported by CSV files are not visible on the Published pages.

Workaround

You must publish the artifacts before they appear as Published and are available to use. If the artifacts are in the imported categories, you might need to add users as collaborators to those categories.

Can't find the artifacts after ZIP import

Issue description

The artifacts that are imported by ZIP files are not visible on the UI but they are available in API calls.

Workaround

It is possible that the wkc-search service was not in a ready state during the ZIP import. To solve the issue, resync with the following command:

curl -X GET "$CPD_URL/v3/glossary_terms/admin/resync?artifact_type=all" -k -H "accept: application/json" -H "Authorization: Bearer $TOKEN"

ZIP Import process creates duplicate artifacts

Issue description

Artifacts in exported ZIP files are identified with artifact identifiers in the Artifact ID column, instead of by the artifact name and category path. However, artifact identifiers for predefined business terms, data classes and classifications differ between each IBM Knowledge Catalog instance. Therefore, imported predefined artifacts are not recognized as being the same as the existing predefined artifacts. Therefore, the import process creates duplicate artifacts.

Workaround

• Delete the existing predefined artifacts from the target IBM Knowledge Catalog instance before you import artifacts. See Deleting published artifacts of the same type.

• For predefined data classes and classifications:

  1. Exclude the [uncategorized] category when you create the ZIP file by exporting artifacts.
  2. Edit the CSV files for data classes and classifications and replace the artifact identifiers for predefined data classes and classifications in the Parent Classification and Parent Data Class columns with the equivalent predefined artifact names and category paths.
  3. Edit the CSV files for other artifacts and replace the artifact identifiers for predefined data classes and classifications with the equivalent predefined artifact names and category paths in the Classifications and Data Classes columns, for any artifacts that have relationships to the predefined artifacts.

  When you import the artifacts, relationships to the exisiting target predefined artifacts are created.

• For predefined business terms:

  1. Exclude the Knowledge Accelerator Sample Personal Data category when you create the ZIP file by exporting artifacts.
  2. Edit the CSV files for business termsa and other artifacts and replace the artifact identifiers for predefined business terms with the equivalent predefined artifact names and category paths in the Related terms column and in any other columns defining business term relationships.

  When you import the artifacts, relationships to the exisiting target predefined artifacts are created.

The predefined reference data sets in the Locations category have the same artifact ID values across all IBM Knowledge Catalog instances, there is no need for a workaround.

Relationships to predefined data classes and classifications imported from ZIP are missing

Issue description

Custom properties and relationship values are not imported.

Workaround

Edit the CSV files and edit the values for the related predefined data classes and classifications from artifact IDs to artifact names and category paths, and then reimport.

Parent topic: Troubleshooting IBM Knowledge Catalog