The Graph Modeling Excel Format

In order to be able to use the Excel sheet for adding custom classes, relations or attributes to concepts, you have to make sure that these prerequisites are met:

You have created a custom scheme for your thesaurus.
The custom scheme is also active for the thesaurus you want to update data for.
For custom class import or update define a type column containing the URI of the class. Details find below.

You have to observe the following rules to create a compliant Excel sheet:

All column headers have to be lower case.
Two areas are differentiated in Graph Modeling's Excel rules:
- the Fixed Area
- the Optional Area
The Fixed Area has to contain certain columns in this order:
1. uri/notation/custom attribute (optional): if present, it must be the first column.
2. scheme (mandatory): contains one or more custom schemes.
3. concept (mandatory): one or more columns containing top concepts or concepts.
The Optional Area is indeed optional, so columns and headers can be present. The following options exist:
- You can add SKOS attributes and relations.
- You also can add custom attributes.
- You can use multiple rows or columns.

Details on the respective columns find below.

Note

In case you have polyhierarchies in your import file, they will be turned into individual concepts.

Check the option Enable Updates if you want to keep them. See: Add Data via Excel

Special Column Headers*

uri: case sensitive (enter URI of a concept into the cells below)
notation: case sensitive (enter custom defined contents of the skos:notation field in the cells below)
<custom attribute>: http://vocabulary.semantic-web.at/cocktail-ontology/image (URI of the attribute, enter value or values into the sheet's cell)
<custom class type>: http://vocabulary.semantic-web.at/cocktail-ontology/Garnish (URI of the class, enter value or values into the sheet's cell)

Fixed Area in the Excel SheetSheet Column: ID column

The ID column (uri / notation / <custom attribute>) is optional.

For an import with the update option this column allows you to define how existing concepts are detected if the default mechanism should not be used. Default is the detection via the preferred label in the concept column.
- The following rules apply for adding an ID column:
  - The ID column has to be the first column in the sheet, if you provide one.
  - There can only be one ID column.
  - For the type <custom attribute> you have to provide the URI of the attribute as a column header.
- You can provide the ID in the following ways:
  - URI: The concept is identified by the URI you provide.
  - Notation: The concept is identified by the value in the skos:notation attribute.
  - <custom attribute>: The concept is identified by the value provided for the custom attribute of the respective custom scheme.
For an export to Excel the URI column including the URIs for the exported concepts is added.
Note
In order to be able to use a custom attribute as a definition in the ID column, you need to have a custom scheme in place that also has been assigned to the project thesaurus in question.

Sheet Column: scheme

The scheme column is optional.

If this column is present, it will create a concept scheme.
The label you provide will be used as a title in the default language of the project for this concept scheme.
Note
If you define this column, you cannot use the import file to create a subtree.

Sheet Column: concept (one is mandatory)

Note

At least one column is required, unless only concept schemes should be created and a scheme column exists.

The label you provide will be used as a preferred label in the default language of the project for this concept.
You can use multiple columns labeled concept to form a hierarchy.

Optional Area in the Excel SheetSheet Column: SKOS attributes

Allowed values: prefLabel@Language, altLabel@Language, hiddenLabel@Language, notation, scopeNote@Language, example@Language, definition@Language
If no language tag is defined, the default language of the project is used.
You can define the language by adding a language tag for example: prefLabel@fr
Note
Preferred labels can only be defined once per language, all other properties can be defined multiple times.

The language codes you can use Graph Modeling should be compliant with the Oracle Java definition.

Sheet Column: SKOS relations

Allowed values: exactMatch, closeMatch, relatedMatch, broaderMatch, related
You can create the relation only if the respective concept already exists.
If no ID column is provided the related concept is identified by the prefLabel in the default language.
If an ID column is available, the related concept is identified by the value in that column.
Note
You have to label the column with the URI of the respective relation.

Important

Custom relations can be imported using Graph Modeling’s tabular import operation, but Graph Modeling tabular export does not support relations of any kind. Specifically, the relation skos:related, SKOS “semantic alignment” relations such as skos:exactMatch, and custom relations are not included in the output of tabular export. Hierarchical relationships (broader/narrower) are represented implicitly in the structure of the exported hierarchy. We are working to support tabular export of relationships in future releases.

Sheet Column: <custom attributes>

Allowed values: custom attributes defined in the corresponding custom scheme.
The custom scheme definitions of the attribute (type, multiple/single) are taken into account.
For language tagged literals the same rules apply as for skos attributes.
Note
You have to label the column with the URI of the respective attribute.

Sheet Column: <custom relations>

Allowed values: custom scheme relations defined on the server.
You can create the relation only if the respective concept already exists.
For identifying the concept the same rules apply as for SKOS relations.
The custom scheme definitions of the relation (type, multiple/single) are taken into account.
Note
You have to label the column with the URI of the respective relation.

Important

Custom relations can be imported using Graph Modeling’s tabular import operation, but Graph Modeling tabular export does not support relations of any kind. Specifically, the relation skos:related, SKOS “semantic alignment” relations such as skos:exactMatch, and custom relations are not included in the output of tabular export. Hierarchical relationships (broader/narrower) are represented implicitly in the structure of the exported hierarchy. We are working to support tabular export of relationships in future releases.

Sheet Column: type

Using the type column you can apply custom classes to concepts on import.
Allowed values: URIs of the custom class to be applied.
You can apply custom classes only to the concept directly. Application on a subtree or concept scheme is not possible.

Note

Empty rows are not allowed.

Column headers are case sensitive.

Use of special characters is restricted inside an Excel sheet.

Graph Modeling will not import, if you use these in a label:

\, ", <, >, tabs and newlines.

Exception: Single quotes ' are allowed.

If you provide an ID column, it defines the way concepts are identified in the import file also when you add data or relations (an ID column contains either URIs, skos:notations or a custom attribute that can be used as ID).

The ID column has to be the first column in the sheet.