Use Regex in Keyword Boosting to Handle Query Inflexion
-
Keyword Tuning is a powerful tool that can be used to change the rank of a document for a query or a related set of queries.
Boost for Query (Precise)
You can be precise with it, so that document X is boosted only for
add checkbox
.In this case, the X may or may not turn up for related queries, such as
adding checkbox
oradd checkboxes
.Boost for Query Forms (Comprehensive)
Alternatively, you can choose to be comprehensive and use a regular expression to boost document X for all the forms of
add checkbox
, so that someone looking foradding checkbox
oradd checkboxes
also gets document X on top.Regex patterns are extremely useful in languages such as Polish and Russian where nouns declensions are aplenty. They are also a useful tool for Spanish, French, and other languages with complex verb conjugations. You can handle these and other inflexions with regular expressions.
SearchUnify Compatible Regular Expression Syntax
Operator Definition Example . Match any character. reali.ze
= realize, realise? Match the preceding character 0 or 1 times. seas?
= sea, seas+ Repeat the preceding character 1 or more times. 123+
= 1233, 12333, 123333...* Repeat the preceding character 0 or more times. 123*
= 123, 1233, 12333, 123333...{} Repetition range for the preceding character. great{1,5}
= great, greatt, greattt, greatttt, greattttt| OR operator realize[s|z]e
= realize, realize() Form a group. extension (0175)|(0172)
= extension 0175, extension 0172[] Match any character or group between the square brackets. ID[0-9]
= ID0, ID1, ID2... ID9- Range. To be used with square brackets. [a-zA-Z]{2,4]
= All words of length 2, 3, or 4.^ Negation. To be used with square brackets. [^0-9]{10}
= Don't match 10-digit numbers, such as 9876323472.Instructions
- Log in to the instance, find Search Tuning, and click Keyword Tuning.
- Select your search client.
- Find the document to be boosted and click Add Keywords.
- Enter a query to boost a document for a keyword or phrase. Or enter a regular expression to boost a document for a query and all its nouns, verbs (tenses), and other variations.
- Click Save.
Official Documentation: Boost Documents for Specific Keywords
-
@kasey-r You're right.
reali.e
will berealiaze
as well. The doc (.
) is a placeholder for any character, not just "s" or "z". The examples in the article are textbook type. In real word, an admin has to be extremely careful. A better way to boost a document for both "realize" and "realise" is to usereali[sz]e
. The square brackets limit the choice to two characters.As for the word order, I'm sure there must be some way. People have written really fat tomes on regular expressions and I don't count myself among those who have read those tomes. A regex expert should be able to tell. From what I know, a (rather clumsy and wordy) way to handle word order for
add checkbox
is to boost the pattern(add[(ed)|(ing)|(s)]? checkbox(es)) | (checkbox(es) add[(ed)|(ing)|(s)]?)
.
-
I'm confused by examples for the . operator and the | operator. Wouldn't reali.ze be read as realiaze, realibze, realicze, realidze...? Also wouldn't realize[s|z]e be read as realizese, realizeze?
Another question: Is it possible to use regex to account for the order of the words in the search? To use your example, would there be a way to also capture "checkbox adding"?
-
This is great to know about. Would be great to use Regular Expressions in the search as well, see idea: https://community.searchunify.com/topic/306/regular-expressions-supported