The data warehouse design is coupled tightly with the structure of the ST36-based XML document. The relational database is the open source PostgreSQL system and makes use of schemas to organize content and organizational structure inside the database. See http://www.postgresql.org for further information on PostgreSQL-specific functionality. The following table lists the active schemas and their roles in the data warehouse.

The following graphic is a tree view of the structure of a complete XML document. Each node represents a container whereby each leaf of the tree represents a table in the data warehouse (xml and work schemas). The naming pattern of tables follows the following rules:

• dashes are replaced with underlines and
• each leaf-node name is prefixed with t_, i.e., publication-reference =t_publication_reference.

The two exceptions to this are the non-XML table t_patent_document_values and the copyright node. The copyright container is stored separately in t_sys_data.

Schemas

Reporting

The reporting schema houses tables critical to data loading, update processes, and queues for optional indexing. Although most tables are for internal use, the table reporting.t_client_load_process contains interesting information for clients around which web status interfaces or triggers can easily be built.

If you have installed the optional Solr index, there is a companion table reporting.t_client_index_process that manages the indexing queue and has the exact structure of the reporting.t_client_load_process.

XML

The XML schema is the heart of the data warehouse and contains tables representing individual containers (sections) of the complete document stored in pure XML.

Table Structure–General

All tables, with the exception of xml.t_patent_document_values, follow the same structure.

The table xml.t_patent_document_values is the master content table containing meta information of the document in standard relational database columns. This table is the basis for the overall container element patent-document.

XML User-Defined Functions

In addition to content columns from which XML content can easily be selected, there are also the following supplemental functions available for common data extraction tasks:

xml.f_patent_document(integer, text)

This function returns an entire XML document given a publication_id as first argument. The second argument is currently unused.

SELECT xml.f_patent_document(1, null);

xml.f_patent_document_s(integer, text)

Similar to the function above, this function returns a SET OF XML documents, one document per row. Arguments are identical to function above.

SELECT xml.f_patent_document_s(t.publication_id, null)
FROM xml.t_patent_document_values AS t
WHERE t.published='20100407';

xml.f_ucid2id(text)

This function takes a ucid (unique character ID) and converts it to an integer publication_id.

SELECT xml.f_patent_document(xml.f_ucid2id('US-5551212-A'), null);