Frequently Asked Questions about Imports, Exports, Extracts, and Loads

What is the difference between an import and a load?

Imports and loads both bring data into Oracle Fusion Cloud Enterprise Data Management, but there are some key differences in how they operate:

Imports

  • Controlled by application registration
  • Used for bound data at the dimension level
  • Require that all bound properties be contained in the import file.
  • File format for the import is controlled by the application type and cannot be changed.
  • Does not support profiles

Loads

  • Focused on viewpoints, rather than dimensions.
  • Can be used for bound and unbound data.
  • Does not require that all properties are contained in the load file.
  • Does not clear out any data chains, and restricts the changes being made to the scope of data as defined by the viewpoint and node set. This means that for hierarchy viewpoints, loads will only affect the structure below the existing top nodes.
  • Supports profiles (see Creating, Copying, and Deleting Loads)

For more differences, see Viewpoint Loads Versus Imports.

When should I use an import versus a load?

Loads and imports are similar in that both load data into dimensions. Which one you select depends on several factors, such as whether you are loading data to bound or unbound viewpoints and properties, what type of load option you require (loads don't support Reset mode), and whether you are loading data from a file or an application connection. Here are some general guidelines for when to use imports versus loads:

  • Use an import when:
    • You are loading bound data from an external system.
    • You are loading data from an application connection.
    • You want to use the Reset mode to clear out existing bound nodes and relationships and then load them from the import file.
    • You want to expand the allowed values for a property.
  • Use a load when:
    • You want to load data to viewpoints or properties regardless of their binding status.
    • You want to load data to a reference or unbound dimension.
    • You want to load to a subset of a hierarchy viewpoint.

What is the difference between Reset, Replace, and Merge modes for imports and loads, and which one should I use?

Table 30-8 Import and Load Modes

Mode Description and Key Features Use Case
Reset (Import Only)

Does a complete reset of bound data including nodes, relationships, and properties.

  • Applies to imports only. Cannot be used with loads.
  • Does not update transaction history
  • Does not trigger subscriptions
  • Impacts partially bound data (such as mapping viewpoints)
Should be used only when a complete reset of the bound data is desired.

Caution: This import mode should be used sparingly, as continuous use will increase the size of your database.

Replace

For imports, does a complete reset of the bound hierarchy set and replaces it with the incoming relationships. Retains the relationship properties for relationships that existed before the import and are in the import file (even for unbound relationship level properties).

For loads, clears out only the relationships in scope for the target viewpoint and node set and replaces them with the incoming relationships. Relationship level property values for relationships that existed before the load and are also in the load are retained (even for unbound relationship level properties).

  • Applies to imports and loads.
  • Does not update transaction history
  • Does not trigger subscriptions
  • Retains nodes and does not impact partially bound data (such as mapping sets).
Recommended for periodic bulk changes from external systems and large changes that do not depend on subscriptions and for which transaction history is not needed.
Merge

Does not clear data sets. Creates a target set of data and compares it to the current data, and then generates a request with the changes needed to convert the current data to match the incoming import or load data.

  • Uses the request process, so transaction history is recorded and subscriptions are triggered.
  • Limited number of changes can be processed per import or load.
  • Enables you to specify additional options for merging data, such as whether to process Deletes and Reorders, whether to auto-submit merge requests, and whether to bypass approvals and commits. See Merge Import Options for imports and Merge Options for loads.
Recommended for periodic bulk changes from external systems and large changes where subscriptions need to be processed and for which transaction history is required.

Additional details about the specifics of how each mode works for imports and loads are available in these topics: