Question

I ve been stuck with this problem for a few weeks now and I have no idea how to solve it. So basically, my index contains data that looks something like this:

[
    { id: 1, value: "some snippet from a paragraph"},
    { id: 1, value: "some different snippet from the same paragraph" }.
    { id: 1, value: "some different snippet from the same paragraph" },
    { id: 2, value: "some snippet from a different paragraph" },
    { id: 2, value: "some different snippet from id no.2 s paragraph " },
    ... and a few more million values
]

如你所知,“一款”分为多层。每个项目都有相同的识别资料,但具有不同的价值。

I used $top and $skip, but the problem with this is that the values in the top n results will contain "duplicates". Like, the top 50 results will contain multiple items that have an ID of 1. What I want to happen is that the top n results should only contain items with unique ID s. Is it possible to just get the item with the highest score of each ID? How can I make this happen?

成果应如此:

[   
    { id: 1, value: "some value" },
    { id: 2, value: "some value" }.
    { id: 3, value: "some value" },
    { id: 4, value: "some value" },
    { id: 5, value: "some value" },
    ... 
]

Answer 1

Have you considered modeling your schema so that each paragraph is a string element in a Collection(Edm.String) field? Each document would look like this:

{
 "id": 1,
 "paragraphs": [
    "some snippet from a paragraph",
    "some different snippet from the same paragraph",
    "some different snippet from the same paragraph"
 ]
}

You can still target the "paragraphs" field with your query, but this will guarantee each top-level documents on appears once in your search results (regardless of how many "paragraph" matched the query).

这里与“支持数据类型”网页链接:https://learn.microsoft.com/en-us/rest/api/searchservice/support-data-types

友情链接