Configure a text index group to search across multiple tables
-
- UpdatedJan 30, 2025
- 5 minutes to read
- Yokohama
- Search Administration
Update tables to the V4 indexing format and include them in a text index group to facilitate searching across multiple tables.
Before you begin
Role required: ts_admin or admin
About this task
Text index groups enable you to index and search across multiple tables.
Performing searches across multiple tables requires normalizing the tables first so the relevance scores mean the same thing for all tables. V4 indexing provides that normalization. Zing uses the Okapi BM25 ranking function to calculate the relevancy score of search results.
The V4 table indexing format is optimized for documentation, such as knowledge base articles. You don't need to reindex tables that use this format when you change the value of an indexing attribute, such as the weight of a field. If you want to use the BM25 ranking function on a single table, you can convert it to the V4 format.
- Configure all of your chosen tables to use the V4 index format if they don't already use it.Note: In new instances, the Catalog Item [sc_cat_item], Knowledge [kb_knowledge], Social Q&A Questions [kb_social_qa_question], and User [sys_user] tables use the V4 index format. If you upgrade your instance from a previous release, the indexing format of these tables remains the same as it was in the older release.
- Add the tables to a text index group so they can be indexed and searched together.Note: You can't add the Task [task] table or any tables that extend it to a text index group.
Procedure
What to do next
You can query the tables in the text index group for string matches using the GlideRecord.addQuery(name, value)
or GlideRecord.addEncodedQuery(name, value)
methods with the
123TEXTINDEXGROUP321
reserved name. For more information about querying text index groups for string matches, see Querying tables in script.
Related Content
- Zing indexes punctuation as part of some words
Zing indexes some punctuation marks as part of some words to improve search results for common search terms.
- Zing indexes some HTML elements
Zing indexes the contents of certain HTML elements to improve document search results. This indexing is addition to the normal indexing by word.
- Configure a table for indexing and searching
Enable Zing text indexing for a table to allow users to search its records for string values. Configure the table's text indexing attributes, stop words, and synonym dictionaries to control search behavior.
- Zing index and search dictionary attributes
The following dictionary attributes either affect the way Zing indexes tables or ranks search results.
- Regenerate a text index for a table
You can regenerate a table text index when you change table stop words or display values.
- Remove an index
You can remove the index for a table if you no longer want the search engine to return results for that table. This procedure also removes the index for all tables that extend the specified table.
- Remove an index for a specific field
You can remove the index for a specific field in a table if you no longer want the search engine to return results for that field.
- Remove the text index for a child table
You can remove the text index for a child table if you no longer want the search engine to return results for that child table.
- Change the query mode of an indexed table
Set the query mode to specify whether searches on an indexed table must match all or only some of the search terms.
- Enable indexing of text in multi-row variable sets
Make text from multi-row Service Catalog variable sets searchable in global search.
- Text indexing statistics and status
To view text indexing statistics and status, navigate to All System Definition Text Indexes .
- Configure tables to use the Japanese tokenizer
Improve search results in Japanese language searches by configuring individual tables to use the Japanese tokenizer.