Stop words
Summarize
Summary of Stop words
Stop words are common terms such as articles, conjunctions, personal pronouns, and prepositions that appear frequently across indexed records but do not add meaningful value to search results. By defining and managing stop words within stop word dictionaries linked to search profiles, ServiceNow AI Search improves search precision and relevancy by removing these terms from user queries.
Show less
Key Features
- Stop word dictionaries: Customers can use default stop word dictionaries or create custom dictionaries linked to search profiles to manage stop words.
- Publishing dictionaries: Published dictionaries activate their stop words for searches using linked profiles; unpublished dictionaries have no effect.
- Multi-word phrases: Stop words can be single words or exact multi-word phrases; phrases are removed only when the query terms match the phrase order and proximity exactly.
- Language and dictionary linkage: Stop word removal is language-dependent; only stop words from dictionaries matching the query language or the default English dictionary are considered.
- Case and punctuation sensitivity: Optional settings enable case-sensitive or punctuation-sensitive matching for stop words within a dictionary.
- Dictionary term-length limit: By default, stop word terms longer than 256 characters are ignored, but this limit can be adjusted with a system property (with caution due to memory impacts).
- Interaction with other search features: Stop words are evaluated in context with Genius Results, result improvement rules, synonyms, typo handling, and wildcard operators to ensure accurate and relevant search behavior.
Practical Usage and Management
- Creating and linking dictionaries: ServiceNow customers can create stop word dictionaries, define stop words, and link them to appropriate search profiles to tailor search query processing.
- Publishing and unpublishing: Publishing a dictionary activates its stop words, while dictionaries not linked to any profile can be deleted to maintain system hygiene.
- Stop word eligibility: To be removed from a query, stop words must be active, in a published dictionary linked to the search profile, and comply with term-length limits.
Benefits for ServiceNow Customers
By configuring stop words properly, customers can enhance search precision and relevancy, ensuring users receive more meaningful search results. This reduces noise from irrelevant common terms and improves overall search efficiency and user satisfaction within their ServiceNow instances.
Stop words remove search query terms that don't produce meaningful results. Improve search precision and relevancy by configuring stop word dictionaries and defining stop words.
Stop words are common terms that appear on a large proportion of indexed records. Examples of such terms can include articles, conjunctions, personal pronouns, and prepositions. By defining these terms as stop words, you can improve search precision and result relevancy for users.
You can define stop words in the default stop word dictionary for a search profile, or create your own stop word dictionaries linked to search profiles. Create and publish stop words for each dictionary. Publish the dictionary to make your stop words take effect for searches that use linked search profiles.
You can delete stop word dictionaries that aren't linked to any search profiles.
Stop word usage in search
When a search query includes a term defined as a stop word, AI Search removes the term from the search query.
For example, a stop word might remove the term my from a search query for upgrade my laptop. This modified search query only matches records containing the more relevant terms upgrade and laptop.
Conditions for stop word eligibility
- It must be active.
- The stop word dictionary that includes it must be published.
- Its length doesn't exceed the dictionary term-length limit.
| Condition | Result |
|---|---|
| The search profile specified for the search query links to a published stop word dictionary for the search query's language. | The system only considers stop words from the linked stop word dictionary.
As an example, suppose a search application's search profile links to a French stop word dictionary. When a user searches in French in this application, only stop words from the linked French stop word dictionary are eligible for removal from the search query terms. |
| The search profile specified for the search query doesn't link to a published stop word dictionary for the search query's language. | The system only considers stop words from the default (English) stop word dictionary.
As an example, suppose a search application's search profile has no linked Spanish stop word dictionary. When a user searches in Spanish in this application, only stop words from the default (English) stop word dictionary are eligible for removal from the search query terms. |
For more details on language support in stop word dictionaries, see Language dependence.
Dictionary term-length limit
By default, AI Search ignores stop word terms that exceed 256 characters in length. You can override this default limit by adding the query.maxDictionaryTermLength system property. The integer value of this system property indicates the maximum allowable character length for stop word and synonym dictionary terms.
Multi-word phrases
A stop word can include either a single word or a multi-word phrase. AI Search removes multi-word phrases from the search query when the query terms exactly match the order and proximity of the phrased words in the stop word definition. If a user searches with these words in a different arrangement, no term removal occurs.
For example, the following table describes AI Search stop word removal behavior for a stop word that includes the multi-word phrase retirement account.
| Search query terms | Stop word removal behavior |
|---|---|
| my retirement account | AI Search removes the terms retirement account from the search query. |
| my bank retirement account | AI Search removes the terms retirement account from the search query. |
| my account retirement | AI Search doesn't remove any terms from the search query. |
| my retirement bank account | AI Search doesn't remove any terms from the search query. |
Case sensitivity
By default, AI Search ignores letter case when comparing stop words with search query terms. You can enable case sensitivity for a stop word dictionary. This option affects all stop words defined in the dictionary.
When case sensitivity is enabled, stop words in the dictionary only match search query terms or phrases that include the same combination of uppercase and lowercase letters.
For example, if you define a RAID stop word in a dictionary with case sensitivity enabled, AI Search ignores the search query term RAID but treats the search query term raid normally.
Punctuation sensitivity
By default, AI Search ignores punctuation when comparing stop words with search query terms. You can enable punctuation sensitivity for a stop word dictionary. This option affects all stop words defined in the dictionary.
When punctuation sensitivity is enabled, stop words in the dictionary only match search query terms or phrases that include the same punctuation.
For example, if you define a one-half stop word in a dictionary with punctuation sensitivity enabled, AI Search removes the search query term one-half but doesn't remove the search query term one half.
Language dependence
AI Search supports stop word dictionaries for all languages activated in your instance. For the list of languages you can activate, see Activate a language.
You can only link one stop word dictionary per language to a given search profile.
Interaction with other search features
The following table describes interactions between stop words and other search features.
| Feature | Interaction with stop words |
|---|---|
| Genius Results | If a search query term that matches a Genius Result configuration's Term trigger is defined as a stop word, AI Search evaluates the trigger before removing that stop word from the search query.
When computing Q&A Genius Results for a search query, AI Search ignores stop words in the query. |
| Lemma and Unicode normalization | If a search query term is defined as a stop word, AI Search removes that term without normalizing it. |
| Result improvement rules | If a search query term that matches a result improvement rule's Query trigger is defined as a stop word, AI Search evaluates the rule trigger before removing the stop word from the search query. |
| Synonyms |
If a synonym definition includes a term defined as a stop word, AI Search retains that term when checking whether the synonym expands a search query. For example, suppose you define as and a as stop words and SAAS↔︎software as a service as a synonym.
|
| Typo handling |
If a misspelled search query term is defined as a stop word, AI Search auto-corrects that term normally instead of removing it. AI Search ignores terms defined as stop words when determining an auto-correction for a misspelled search query term. For example, if you define email as a stop word, AI Search doesn't offer email as an auto-correction when you search for emial. |
| Wildcard operators | When expanding search terms that contain % or * wildcard operators, AI Search ignores terms defined as stop words. For example, suppose you define the and their as stop words. A search for the* won't expand to match the or their, but will still match non-stop word terms such as there and these. |
Publishing stop word dictionaries
Publishing a stop word dictionary makes its settings and active stop words take effect in all linked search profiles.
Publishing a search profile updates the profile's stop word dictionary links but doesn't update the dictionary's settings or active stop words in AI Search.