分組搜尋

分組搜尋可讓 Milvus 根據指定欄位的值將搜尋結果分組，以便在更高層級彙整資料。 例如，您可以使用基本的 ANN 搜尋來尋找與當前書籍相似的書籍，但若要找出可能涉及該書所討論主題的書籍類別，則可使用分組搜尋。本主題將說明如何使用分組搜尋，並列出相關的重要考量事項。

概述

當搜尋結果中的實體在某個標量欄位中具有相同的值時，這表示它們在特定屬性上相似，這可能會對搜尋結果產生負面影響。

假設某個集合儲存了多個文件（以docId 表示）。 為了在將文件轉換為向量時盡可能保留語義資訊，每個文件都會被分割成較小且易於管理的段落（或片段），並作為獨立實體儲存。即使文件已被分割成較小的部分，使用者通常仍希望識別出哪些文件最符合其需求。

Ann 搜尋

當對此類文獻集執行近似最近鄰（ANN）搜尋時，搜尋結果可能包含來自同一份文獻的數個段落，這可能導致其他文獻被忽略，而這可能不符合預期的使用情境。

分組搜尋

為提高搜尋結果的多樣性，您可在搜尋請求中加入group_by_field 參數以啟用分組搜尋。如圖所示，您可以將group_by_field 設定為docId 。收到此請求後，Milvus將：

• 根據提供的查詢向量執行人工神經網路 (ANN) 搜尋，以找出與查詢最相似的所有實體。

• 根據指定的「group_by_field 」（例如docId ）將搜尋結果進行分組。

• 根據limit 參數的定義，針對每個群組返回頂部結果，並包含該群組中最相似的實體。

執行分組搜尋

本節提供範例程式碼，用以示範分組搜尋的使用方法。以下範例假設集合包含以下欄位：id 、vector 、chunk 以及docId 。

[
        {"id": 0, "vector": [0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592], "chunk": "pink_8682", "docId": 1},
        {"id": 1, "vector": [0.19886812562848388, 0.06023560599112088, 0.6976963061752597, 0.2614474506242501, 0.838729485096104], "chunk": "red_7025", "docId": 5},
        {"id": 2, "vector": [0.43742130801983836, -0.5597502546264526, 0.6457887650909682, 0.7894058910881185, 0.20785793220625592], "chunk": "orange_6781", "docId": 2},
        {"id": 3, "vector": [0.3172005263489739, 0.9719044792798428, -0.36981146090600725, -0.4860894583077995, 0.95791889146345], "chunk": "pink_9298", "docId": 3},
        {"id": 4, "vector": [0.4452349528804562, -0.8757026943054742, 0.8220779437047674, 0.46406290649483184, 0.30337481143159106], "chunk": "red_4794", "docId": 3},
        {"id": 5, "vector": [0.985825131989184, -0.8144651566660419, 0.6299267002202009, 0.1206906911183383, -0.1446277761879955], "chunk": "yellow_4222", "docId": 4},
        {"id": 6, "vector": [0.8371977790571115, -0.015764369584852833, -0.31062937026679327, -0.562666951622192, -0.8984947637863987], "chunk": "red_9392", "docId": 1},
        {"id": 7, "vector": [-0.33445148015177995, -0.2567135004164067, 0.8987539745369246, 0.9402995886420709, 0.5378064918413052], "chunk": "grey_8510", "docId": 2},
        {"id": 8, "vector": [0.39524717779832685, 0.4000257286739164, -0.5890507376891594, -0.8650502298996872, -0.6140360785406336], "chunk": "white_9381", "docId": 5},
        {"id": 9, "vector": [0.5718280481994695, 0.24070317428066512, -0.3737913482606834, -0.06726932177492717, -0.6980531615588608], "chunk": "purple_4976", "docId": 3},
]

在搜尋請求中，將group_by_field 和output_fields 兩者皆設定為docId 。Milvus 會依據指定的欄位對結果進行分組，並從每個分組中返回最相似的實體，同時包含每個返回實體的docId 值。

from pymilvus import MilvusClient

client = MilvusClient(
    uri="http://localhost:19530",
    token="root:Milvus"
)

query_vectors = [
    [0.14529211512077012, 0.9147257273453546, 0.7965055218724449, 0.7009258593102812, 0.5605206522382088]]

# Group search results
res = client.search(
    collection_name="my_collection",
    data=query_vectors,
    limit=3,
    group_by_field="docId",
    output_fields=["docId"]
)

# Retrieve the values in the `docId` column
doc_ids = [result['entity']['docId'] for result in res[0]]

import io.milvus.v2.client.ConnectConfig;
import io.milvus.v2.client.MilvusClientV2;
import io.milvus.v2.service.vector.request.SearchReq
import io.milvus.v2.service.vector.request.data.FloatVec;
import io.milvus.v2.service.vector.response.SearchResp

MilvusClientV2 client = new MilvusClientV2(ConnectConfig.builder()
        .uri("http://localhost:19530")
        .token("root:Milvus")
        .build());

FloatVec queryVector = new FloatVec(new float[]{0.14529211512077012f, 0.9147257273453546f, 0.7965055218724449f, 0.7009258593102812f, 0.5605206522382088f});
SearchReq searchReq = SearchReq.builder()
        .collectionName("my_collection")
        .data(Collections.singletonList(queryVector))
        .topK(3)
        .groupByFieldName("docId")
        .outputFields(Collections.singletonList("docId"))
        .build();

SearchResp searchResp = client.search(searchReq);

List<List<SearchResp.SearchResult>> searchResults = searchResp.getSearchResults();
for (List<SearchResp.SearchResult> results : searchResults) {
    System.out.println("TopK results:");
    for (SearchResp.SearchResult result : results) {
        System.out.println(result);
    }
}

// Output
// TopK results:
// SearchResp.SearchResult(entity={docId=5}, score=0.74767184, id=1)
// SearchResp.SearchResult(entity={docId=2}, score=0.6254269, id=7)
// SearchResp.SearchResult(entity={docId=3}, score=0.3611898, id=3)

import (
    "context"
    "fmt"

    "github.com/milvus-io/milvus/client/v2/entity"
    "github.com/milvus-io/milvus/client/v2/milvusclient"
)

ctx, cancel := context.WithCancel(context.Background())
defer cancel()

milvusAddr := "localhost:19530"
client, err := milvusclient.New(ctx, &milvusclient.ClientConfig{
    Address: milvusAddr,
})
if err != nil {
    fmt.Println(err.Error())
    // handle error
}
defer client.Close(ctx)

queryVector := []float32{0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592}

resultSets, err := client.Search(ctx, milvusclient.NewSearchOption(
    "my_collection", // collectionName
    3,               // limit
    []entity.Vector{entity.FloatVector(queryVector)},
).WithANNSField("vector").
    WithGroupByField("docId").
    WithOutputFields("docId"))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}

for _, resultSet := range resultSets {
    fmt.Println("IDs: ", resultSet.IDs.FieldData().GetScalars())
    fmt.Println("Scores: ", resultSet.Scores)
    fmt.Println("docId: ", resultSet.GetColumn("docId").FieldData().GetScalars())
}

import { MilvusClient, DataType } from "@zilliz/milvus2-sdk-node";

const address = "http://localhost:19530";
const token = "root:Milvus";
const client = new MilvusClient({address, token});

var query_vector = [0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592]

res = await client.search({
    collection_name: "my_collection",
    data: [query_vector],
    limit: 3,
    group_by_field: "docId"
})

// Retrieve the values in the `docId` column
var docIds = res.results.map(result => result.entity.docId)

export CLUSTER_ENDPOINT="http://localhost:19530"
export TOKEN="root:Milvus"

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/search" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
    "collectionName": "my_collection",
    "data": [
        [0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592]
    ],
    "annsField": "vector",
    "limit": 3,
    "groupingField": "docId",
    "outputFields": ["docId"]
}'

#include "milvus/MilvusClientV2.h"
#include <iostream>
#include <stdexcept>
#include <vector>

auto client = milvus::MilvusClientV2::Create();

milvus::ConnectParam connect_param{"http://localhost:19530", "root:Milvus"};
auto status = client->Connect(connect_param);
if (!status.IsOk()) {
    throw std::runtime_error(status.Message());
}

std::vector<float> query_vector = {0.3580376395471989f, -0.6023495712049978f, 0.18414012509913835f, -0.26286205330961354f, 0.9029438446296592f};
auto request = milvus::SearchRequest()
                   .WithCollectionName("my_collection")
                   .AddFloatVector(query_vector)
                   .WithLimit(3)
                   .WithAnnsField("vector")
                   .WithGroupByField("docId")
                   .AddOutputField("docId");

milvus::SearchResponse response;
status = client->Search(request, response);
if (!status.IsOk()) {
    throw std::runtime_error(status.Message());
}

for (auto& result : response.Results().Results()) {
    std::cout << "TopK results:" << std::endl;
    milvus::EntityRows output_rows;
    status = result.OutputRows(output_rows);
    if (!status.IsOk()) {
        throw std::runtime_error(status.Message());
    }
    for (const auto& row : output_rows) {
        std::cout << "\t" << row << std::endl;
    }
}

在上述請求中，limit=3 表示系統將返回來自三個群組的搜尋結果，每個群組包含一個與查詢向量最相似的實體。

設定群組大小

預設情況下，分組搜尋每組僅返回一個實體。若需每組返回多個結果，請調整group_size 和strict_group_size 參數。

# Group search results

res = client.search(
    collection_name="my_collection", 
    data=query_vectors, # query vector
    limit=5, # number of groups to return
    group_by_field="docId", # grouping field
    group_size=2, # p to 2 entities to return from each group
    strict_group_size=True, # return exact 2 entities from each group
    output_fields=["docId"]
)

FloatVec queryVector = new FloatVec(new float[]{0.14529211512077012f, 0.9147257273453546f, 0.7965055218724449f, 0.7009258593102812f, 0.5605206522382088f});
SearchReq searchReq = SearchReq.builder()
        .collectionName("my_collection")
        .data(Collections.singletonList(queryVector))
        .topK(5)
        .groupByFieldName("docId")
        .groupSize(2)
        .strictGroupSize(true)
        .outputFields(Collections.singletonList("docId"))
        .build();

SearchResp searchResp = client.search(searchReq);

List<List<SearchResp.SearchResult>> searchResults = searchResp.getSearchResults();
for (List<SearchResp.SearchResult> results : searchResults) {
    System.out.println("TopK results:");
    for (SearchResp.SearchResult result : results) {
        System.out.println(result);
    }
}

// Output
// TopK results:
// SearchResp.SearchResult(entity={docId=5}, score=0.74767184, id=1)
// SearchResp.SearchResult(entity={docId=5}, score=-0.49148706, id=8)
// SearchResp.SearchResult(entity={docId=2}, score=0.6254269, id=7)
// SearchResp.SearchResult(entity={docId=2}, score=0.38515577, id=2)
// SearchResp.SearchResult(entity={docId=3}, score=0.3611898, id=3)
// SearchResp.SearchResult(entity={docId=3}, score=0.19556211, id=4)

import (
    "context"
    "fmt"

    "github.com/milvus-io/milvus/client/v2/entity"
    "github.com/milvus-io/milvus/client/v2/milvusclient"
)

ctx, cancel := context.WithCancel(context.Background())
defer cancel()

milvusAddr := "localhost:19530"
client, err := milvusclient.New(ctx, &milvusclient.ClientConfig{
    Address: milvusAddr,
})
if err != nil {
    fmt.Println(err.Error())
    // handle error
}
defer client.Close(ctx)

queryVector := []float32{0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592}

resultSets, err := client.Search(ctx, milvusclient.NewSearchOption(
    "my_collection", // collectionName
    5,               // limit
    []entity.Vector{entity.FloatVector(queryVector)},
).WithANNSField("vector").
    WithGroupByField("docId").
    WithStrictGroupSize(true).
    WithGroupSize(2).
    WithOutputFields("docId"))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}

for _, resultSet := range resultSets {
    fmt.Println("IDs: ", resultSet.IDs.FieldData().GetScalars())
    fmt.Println("Scores: ", resultSet.Scores)
    fmt.Println("docId: ", resultSet.GetColumn("docId").FieldData().GetScalars())
}

import { MilvusClient, DataType } from "@zilliz/milvus2-sdk-node";

const address = "http://localhost:19530";
const token = "root:Milvus";
const client = new MilvusClient({address, token});

var query_vector = [0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592]

res = await client.search({
    collection_name: "my_collection",
    data: [query_vector],
    limit: 5,
    group_by_field: "docId",
    group_size: 2,
    strict_group_size: true
})

// Retrieve the values in the `docId` column
var docIds = res.results.map(result => result.entity.docId)

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/search" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
    "collectionName": "my_collection",
    "data": [
        [0.3580376395471989, -0.6023495712049978, 0.18414012509913835, -0.26286205330961354, 0.9029438446296592]
    ],
    "annsField": "vector",
    "limit": 5,
    "groupingField": "docId",
    "groupSize":2,
    "strictGroupSize":true,
    "outputFields": ["docId"]
}'

#include "milvus/MilvusClientV2.h"
#include <iostream>
#include <stdexcept>
#include <vector>

auto client = milvus::MilvusClientV2::Create();

milvus::ConnectParam connect_param{"http://localhost:19530", "root:Milvus"};
auto status = client->Connect(connect_param);
if (!status.IsOk()) {
    throw std::runtime_error(status.Message());
}

std::vector<float> query_vector = {0.3580376395471989f, -0.6023495712049978f, 0.18414012509913835f, -0.26286205330961354f, 0.9029438446296592f};
auto request = milvus::SearchRequest()
                   .WithCollectionName("my_collection")
                   .AddFloatVector(query_vector)
                   .WithLimit(5)
                   .WithAnnsField("vector")
                   .WithGroupByField("docId")
                   .WithGroupSize(2)
                   .WithStrictGroupSize(true)
                   .AddOutputField("docId");

milvus::SearchResponse response;
status = client->Search(request, response);
if (!status.IsOk()) {
    throw std::runtime_error(status.Message());
}

for (auto& result : response.Results().Results()) {
    std::cout << "TopK results:" << std::endl;
    milvus::EntityRows output_rows;
    status = result.OutputRows(output_rows);
    if (!status.IsOk()) {
        throw std::runtime_error(status.Message());
    }
    for (const auto& row : output_rows) {
        std::cout << "\t" << row << std::endl;
    }
}

在上述範例中：

• group_size: 指定每組期望回傳的實體數量。例如，設定group_size=2 表示每個群組（或每個docId ）理想上應回傳兩個最相似的段落（或區塊）。若未設定group_size ，系統預設會每組回傳一個結果。

• strict_group_size: 此布林參數控制系統是否應嚴格執行由group_size 設定的數量。當strict_group_size=True 時，系統將嘗試在每個群組中包含由group_size 指定的確切實體數量（例如兩個段落），除非該群組中的資料不足。 預設情況下（strict_group_size=False ），系統會優先確保符合limit 參數所指定的群組數量，而非確保每個群組包含group_size 個實體。在資料分布不均勻的情況下，此方法通常更為高效。

有關參數的更多詳細資訊，請參閱search。

依標量欄位排序群組Compatible with Milvus 3.0.x

您可以將「分組搜尋」(Grouping Search) 與「分組搜尋結果」(order_by_fields ) 結合使用，以標量欄位對群組進行排序。當您希望各群組的結果各不相同，但仍希望群組遵循價格或評分等與業務相關的排序順序時，此方法非常有用。

以下範例會依據「category 」對搜尋結果進行分組，每組最多返回三個實體，並依據「price 」將返回的分組從低到高排序。

res = client.search(
    collection_name="product_catalog",
    data=query_vectors,
    anns_field="embedding",
    limit=20,
    group_by_field="category",
    group_size=3,
    strict_group_size=True,
    output_fields=["category", "price", "rating"],
    order_by_fields=[
        {"field": "price", "order": "asc"}
    ],
)

import io.milvus.v2.service.vector.request.SearchReq;
import io.milvus.v2.service.vector.request.data.FloatVec;
import io.milvus.v2.service.vector.request.aggregation.AggDirection;
import io.milvus.v2.service.vector.request.aggregation.OrderByField;
import io.milvus.v2.service.vector.response.SearchResp;
import java.util.List;

// Prerequisite: client is connected to Milvus and product_catalog is loaded.
FloatVec queryVector = new FloatVec(new float[]{0.14529211512077012f, 0.9147257273453546f, 0.7965055218724449f, 0.7009258593102812f, 0.5605206522382088f});
SearchReq request = SearchReq.builder()
    .collectionName("product_catalog")
    .data(List.of(queryVector))
    .annsField("embedding")
    .topK(20)
    .groupByFieldName("category")
    .groupSize(3)
    .strictGroupSize(true)
    .outputFields(List.of("category", "price", "rating"))
    .orderByFields(List.of(OrderByField.builder()
        .fieldName("price").direction(AggDirection.ASC).build()))
    .build();
SearchResp response = client.search(request);
System.out.println(response.getSearchResults());

// Prerequisite: client is connected to Milvus and product_catalog is loaded.
const queryVector = [0.14529211512077012, 0.9147257273453546, 0.7965055218724449, 0.7009258593102812, 0.5605206522382088];
const response = await client.search({
  collection_name: "product_catalog",
  data: [queryVector],
  anns_field: "embedding",
  limit: 20,
  group_by_field: "category",
  group_size: 3,
  strict_group_size: true,
  output_fields: ["category", "price", "rating"],
  order_by_fields: [{ field: "price", order: "asc" }],
});
console.log(response.results);

import (
    "fmt"
    "github.com/milvus-io/milvus/client/v3/entity"
    "github.com/milvus-io/milvus/client/v3/milvusclient"
)

// Prerequisite: client is connected to Milvus and product_catalog is loaded.
queryVector := []float32{0.14529211512077012, 0.9147257273453546, 0.7965055218724449, 0.7009258593102812, 0.5605206522382088}
results, err := client.Search(ctx, milvusclient.NewSearchOption(
    "product_catalog", 20, []entity.Vector{entity.FloatVector(queryVector)},
).
    WithANNSField("embedding").
    WithGroupByField("category").
    WithGroupSize(3).
    WithStrictGroupSize(true).
    WithOutputFields("category", "price", "rating").
    WithSearchParam("order_by_fields", "price:asc"))
if err != nil {
    panic(err)
}
for _, result := range results {
    fmt.Println(result.IDs, result.Scores)
    fmt.Println(result.GetColumn("category"), result.GetColumn("price"), result.GetColumn("rating"))
}

# Prerequisite: set CLUSTER_ENDPOINT and TOKEN for your Milvus instance.
curl --request POST \
  --url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/search" \
  --header "Authorization: Bearer ${TOKEN}" \
  --header "Content-Type: application/json" \
  --data '{
    "collectionName": "product_catalog",
    "data": [[0.14529211512077012, 0.9147257273453546, 0.7965055218724449, 0.7009258593102812, 0.5605206522382088]],
    "annsField": "embedding",
    "limit": 20,
    "groupingField": "category",
    "groupSize": 3,
    "strictGroupSize": true,
    "outputFields": ["category", "price", "rating"],
    "orderByFields": ["price:asc"]
  }'

#include "milvus/MilvusClientV2.h"
#include <iostream>
#include <stdexcept>

// Prerequisite: client is connected to Milvus and product_catalog is loaded.
std::vector<float> query_vector = {0.14529211512077012f, 0.9147257273453546f, 0.7965055218724449f, 0.7009258593102812f, 0.5605206522382088f};
auto request = milvus::SearchRequest()
    .WithCollectionName("product_catalog")
    .AddFloatVector(query_vector)
    .WithAnnsField("embedding")
    .WithLimit(20)
    .WithGroupByField("category")
    .WithGroupSize(3)
    .WithStrictGroupSize(true)
    .AddOutputField("category")
    .AddOutputField("price")
    .AddOutputField("rating")
    .AddOrderByField(milvus::OrderByField("price", milvus::AggregationDirection::ASC));
milvus::SearchResponse response;
auto status = client->Search(request, response);
if (!status.IsOk()) { throw std::runtime_error(status.Message()); }
for (const auto& result : response.Results().Results()) {
    milvus::EntityRows rows;
    status = result.OutputRows(rows);
    if (!status.IsOk()) { throw std::runtime_error(status.Message()); }
    std::cout << rows << std::endl;
}

在上述請求中，limit=20 表示 Milvus 會選取最多 20 個群組，而非 20 個實體。由於group_size=3 ，因此平鋪的結果清單總計最多可包含 60 個實體。

當您將order_by_fields 與group_by_field 搭配使用時，Milvus 會根據每個群組中首個實體的指定標量欄位值來排序群組。在每個群組內部，實體仍會依據其與查詢向量的相似度分數進行排序。

注意事項

• 索引設定： 此分組功能僅適用於使用以下索引類型的集合：FLAT、IVF_FLAT、IVF_SQ8、HNSW、HNSW_PQ、HNSW_PRQ、HNSW_SQ、DISKANN、SPARSE_INVERTED_INDEX。

• 群組數量：limit 參數用以控制搜尋結果所來自的群組數量，而非各群組內的具體實體數量。設定適當的limit 有助於控制搜尋多樣性與查詢效能。若資料分佈密集或需考量效能問題，降低limit 可減少運算成本。

• 每組實體數：group_size 參數控制每組返回的實體數量。根據您的使用情境調整group_size ，可增加搜尋結果的豐富度。然而，若資料分佈不均，某些組別返回的實體數可能會少於group_size 所指定的數量，特別是在資料有限的情境下。

• 嚴格的分組大小：當設定為 `strict_group_size=True` 時，系統將嘗試為每個分組返回指定數量的實體（group_size ），除非該分組中的資料不足。此設定可確保每個分組的實體數量一致，但在資料分佈不均或資源有限的情況下，可能會導致效能下降。若無需嚴格控制實體數量，設定 `strict_group_size=False ` 可提升查詢速度。

• 若查詢向量已存在於目標集合中，請考慮使用ids ，而非在搜尋前先檢索這些向量。詳細資訊請參閱《主鍵搜尋》。