Index areas and full-text indexes

An index area is a file system directory that contains the files for the full-text indexes that belong to the area. You create index areas manually. An object store can have one or more index areas.

Objects that belong to the same full-text index must share the following characteristics:

• Belong to the same indexable base class or to some subclass of the same indexable base class.
• Have the same partition property value if a string partition property is configured for the object store.
• Have a partition property value within the same date range if a date partition property is configured for the object store.

Content Platform Engine automatically creates full-text indexes to satisfy those requirements. Also, indexes might be created because the existing indexes are full as defined by the configured capacity for the index area.

For each indexed object, one index entry exists in one full-text index. This entry contains information for the content elements and also any CBR-enabled properties of the object. A full-text search returns a maximum of one search result for any object that matches the search criteria.

Virus scanning for index areas

It is not recommended to run virus scans on index areas as it can remove segments or files from the full-text index. These changes can alter the full-text index sizes and prevent content retrieval by text search services like IBM Content Search Services or ElasticSearch. In such cases, the full-text index is corrupted and re-indexing is often required.