Spell Check

When a query returns no results, the Spell Check feature locates similar-looking words in the index (Data Driven) or in a custom list (Dictionary Driven). The Spell Check feature automatically runs a new query based on the similar word. See Spell Check vs. Auto-Suggest for more perspective.

Spell Check Is Triggered by No Results

The Spell Check feature reacts to No Results events, as noted on our Spell Check vs. Auto-Suggest page.

Spell Check doesn’t change every misspelled query. It runs only when the original search returns no results. If a misspelled query already returns results, Site Search shows those results instead of rerunning the query with a corrected term.

The following image shows a spell check response to a “no results” situation:

Search results highlight the misspelling

Spell Check must be enabled by a checkbox. Be sure to Publish your changes.

Dictionary-Driven Spell Check

If dictionary-driven, Site Search uses a dictionary that you provide. 

A spell check configuration panel showing the enable checkbox, method dropdown with

You can build the dictionary one word at a time or upload a .txt file. Use this option for product names, acronyms, industry terms, and other correct one-word entries that Site Search should use as spell-check targets.

TXT Upload Format

Prepare the upload file as plain text. Put one dictionary term on each line. Don’t include a header row or commas. Save the file with a .txt extension.

catalog
invoice
dashboard
enrollment
  • Use one dictionary word per line.
  • Don't use commas to separate terms on the same line.
  • Duplicate entries are ignored.
  • The maximum file size is 10 MB.
  • Site Search can accommodate up to 25,000 spell-check entries.

When you upload a file, Site Search appends the uploaded terms to the existing dictionary list. Review the list after upload, then click Save Draft and Publish to make the dictionary active.

Test Dictionary-Driven Spell Check

To test dictionary-driven Spell Check, search for a deliberate misspelling of one of your dictionary terms in a way that returns no results. For example, after adding catalog, test a query such as catlog. If that misspelled query already returns results, Spell Check won't rerun it.

Data-Driven Spell Check

You can direct the Spell Check feature to look for keywords in one of the search fields. Heading fields, for instance, often contain a dense set of high-relevance keywords. Select the Data-Driven spell check method. This opens a new control for selecting a search field.

Data-driven spellcheck also depends on the spellcheck source field in your app schema (for example, text_spellcheck). The field is auto-created for new Site Search apps, but copied source content can vary by connector.

Note: For custom apps, title and content are copied into text_spellcheck by default. For Drupal and Sitecore implementations, fields may not be copied initially, so you may need to add copy-field rules so the spellcheck source field receives content from your chosen fields.

Note: In multilingual apps, spellcheck configuration and dictionary files are language-specific. Configure and validate spellcheck for each active language.

Data+Dictionary Driven

Site Search supports using both Spell Check features at the same time. You can draw spelling matches from the configured data source and supplement this resource by adding individual words to a dictionary.

Troubleshooting Data-Driven Spell Check

Use these checks if data-driven spellcheck is enabled but suggestions are missing or low quality:

  • Verify the spellcheck source field is receiving content from the fields you expect.
  • After schema or copy-field changes, run a full re-index so existing documents populate the updated field.
  • For Drupal migrations, see Upload Configurations and Reducing Nested Clauses with Copy Fields for schema update patterns.

For API behavior details, see Search API.

Articles in this section