Enabling spelling correction and updating spelling dictionaries

The updateSpellingDictionaries operation of the Data Ingest Web Service performs two functions at once: when it runs, it enables spelling correction and DYM features, and also rebuilds the spelling dictionaries for this data domain.

The operation allows you to rebuild the spelling dictionaries for spelling correction from the data corpus while continuing to issue queries and updates to the Oracle Endeca Server, and without stopping and restarting the Dgraph process. Run this operation after you have added data records to the data domain, to enable spelling correction in the Dgraph process for this data domain.

Important: In a cluster of Dgraphs serving a specific data domain (also known as the data domain cluster), this operation runs successfully on the Dgraph node that can accept updating requests for this data domain (this is the leader Dgraph node). The Endeca Server automatically routes this operation to the leader node in the data domain cluster.

During the data ingest process, you can run the updateSpellingDictionaries operation periodically to update the spelling dictionary used by the Dgraph for Automatic Spelling Correction and DYM.

The updateSpellingDictionaries operation performs the following actions:

Crawls the text search index for all terms which meet the constraint settings.
The constraint settings include minimum word occurrences and maximum and minimum number of characters, for records and attribute values. The Dgraph uses these constraints to update the spelling dictionary. You can change them in the Global Configuration Record.
Updates the spelling dictionaries stored in the Dgraph for processing of all queries arriving after this update to the data files. The Dgraph uses these updated dictionaries when processing all future queries.

The Dgraph applies the updated settings while continuing to run queries and without needing to restart.

For information on how to run the updateSpellingDictionaries operation directly on the Endeca Server, see the Oracle Endeca Server Data Loading Guide, or, if you use Integrator ETL for data loading, see its documentation.