The extraction process consists of getting a set of consistent data from a source and create an archive stored in a H2 database file in the Extract Server, while guaranteeing the referential integrity of the database. All of the extractions defined in a request are stored in the same archive and are based on the same data model. The definition of the referenced data to retrieve is defined in the extraction package.

The extraction, based on the source data model, retrieves all of the dependent data so as to get a consistent set of data in the archive. Filters are applied during the extraction to restrict the rows selected, avoiding retrieving all rows from the tables.

The results of the extraction are kept for each extraction instance, unless the instance is deleted. The extraction results can also be injected as many times as necessary, as long as the target database is compatible, meaning that it contains the same table names and structures as the extraction tables stored in the archive.

For each extraction process, the extraction request contains the extraction branches, each of it containing extractions. The extraction request is made of one or multiple extraction definition. The initial extraction is the first request made to the database to get initial data from one table. Each referenced or dependent data linked to this first set of data fetched at level 0 constitute the next level, and so on and so forth, and eventually, each fetched data has a new level. As a level of extraction is represented by a table, filters can be applied on levels.


Extractions are accessed and managed in the Extractions view.