The Search Service performs each search on a search index instead of the Model repository or profiling warehouse. The search index enables faster searches and searches on content from the Model repository and profiling warehouse.
The Search Service generates the search index based on content in the Model repository and profiling warehouse. The Search Service contains extractors to extract content from each repository.
The Search Service contains the following extractors:
Model Repository extractor
Extracts content from a Model repository.
Business Glossary extractor
Extracts business glossary terms from the Model repository.
Profiling Warehouse extractor
Extracts column profiling results and domain discovery results from a profiling warehouse.
The Search Service indexes all content that it extracts. The Search Service maintains one search index for all extracted content. If a search index does not exist when the Search Service starts, the Search Service generates the search index.
During the initial extraction, the Search Service extracts and indexes all content. After the first extraction, the Search Service updates the search index based on content that has been added to, changed in, or removed from the Model repository and profiling warehouse since the previous extraction. You can configure the interval at which the Search Service generates the search index.
The Search Service extracts and indexes batches of objects. If it fails to extract or index an object, it retries again. After the third attempt, the Search Service ignores the object, writes an error message to the Search Service log, and then processes the next object.
The Search Service stores the search index in files in the extraction directory that you specify when you create the Search Service.