
Announcements
Does anyone know what is the range of the Dataverse search score returned when calling the "search rows (preview)" action in Power Automate? When searching, we get a JSON result similar to the below sample. I have seen scores that are both positive and negative but I can't tell which have a high confidence level. I haven't been able to find any documentation around this. I'd like to use this score to filter results that don't have a high matching probability (or confidence score). It would have been nice if this was a 0 to 1 range or 0-100. But it's not.
thanks!
{
"@search.score": -26.41175079345703,
"@search.highlights": {
"description": [
"{crmhit}myKeywordHere{/crmhit} "
]
},
"@search.entityname": "myEntity123",
"@search.objectid": "3404b2a5-72d0-ee11-9079-000d3a1b97b9",
x
x
x
x}
It's using BM25 - https://learn.microsoft.com/en-us/azure/search/index-similarity-and-scoring#scoring-algorithms-used-in-full-text-search
There's no maximum in this algorithm as everything is calculated using various factors.
You can request moving back to "Classic" as specified in the doc for scores out of max of 1 (or 100 percent)