The following examples illustrate the effects on the search results of the crossfieldMatch and partialMatch attributes.
cross-field matching example
Suppose that a shopper searches for matches on the string “men’s shoes” through a search interface that is configured as follows:
The
crossFieldMatchattribute of the search interface is set to Always, as follows:"crossFieldMatch": "always"The members of the search interface include a dimension named
product.description, as follows:"fields": [
{
"attribute": "product.category "
},
{
"attribute": "product.description"
}
]
The search results for a search on the string “men’s shoes” will include the following records:
Records tagged to dimension values in the dimension
product.categorythat match both “men’s” and “shoes”.Records with a property named
product.descriptionthat matches both “men’s” and “shoes”.Records tagged to one or more dimension values in the dimension
product.categorythat match “men’s” and that have a property namedproduct.descriptionwhose value matches “shoes”.Records tagged to one or more dimension values in the dimension
product.categorythat match “shoes” and that have a property namedproduct.descriptionwhose value matches “men’s”.
partial match example
Suppose that a shopper searches for matches on the string “men’s large blue suede shoes” through a search interface that is configured with the following partialMatch values:
"partialMatch":
{"maxWordsOmitted": 1,
"minWordsIncluded": 2
}
The value of maxWordsOmitted specifies that a property or dimension value cannot match the shopper’s search string if it omits more than one word of the search string. Thus, only a property or dimension value that includes at least four of the five words in the search string “men’s large blue suede shoes” will match it.
The value of minWordsIncluded specifies that a property or dimension value cannot match any search string if it includes fewer than two words of the search string.
Note: Setting minWordsIncluded to a low value can increase the possibility of getting irrelevant results. Suppose, for example, that minWordsIncluded is set to 1.
If the shopper enters the search string “purple jeans”, the search results include all records that match “purple jeans”. If it finds no records that match “purple jeans”, it returns any records that match “purple” or “jeans” – neither of which results is relevant to the shopper’s search.
To eliminate from the search results any records that match only the single words “purple “ or “jeans”, set minWordsIncluded to 2. The search results now include only records that match “purple jeans”; if no records match “purple jeans” there will be no records in the search results.
