數字欄位

數字欄位是儲存數值的標量欄位。這些數值可以是整數（整數）或小數（浮點數）。它們通常用來表示數量、測量值或任何需要以數學方式處理的資料。

下表描述 Milvus 中可用的數字欄位的資料類型。

欄位類型	說明
BOOL	布林類型，用於儲存true 或false ，適合描述二進位狀態。
INT8	8 位整數，適合儲存小範圍的整數資料。
INT16	16 位元整數，適用於中範圍整數資料。
INT32	32 位元整數，適合儲存一般整數資料，例如產品數量或使用者 ID。
INT64	64 位元整數，適合儲存時間戳記或識別碼等大範圍資料。
FLOAT	32 位元浮點數，適用於需要一般精確度的資料，例如等級或溫度。
DOUBLE	64 位元雙精度浮點數，適用於高精度資料，例如財務資訊或科學計算。

若要宣告數字欄位，只要將datatype 設定為可用的數值資料類型之一即可。例如，DataType.INT64 表示整數欄位，DataType.FLOAT 表示浮點欄位。

新增數字欄位

若要儲存數值資料，請在集合模式中定義一個數字欄位。以下是一個有兩個數字欄位的集合模式範例：

• age：儲存整數資料，允許空值，預設值為18 。

• price：儲存浮點數資料，允許空值，但沒有預設值。

# Import necessary libraries
from pymilvus import MilvusClient, DataType

# Define server address
SERVER_ADDR = "http://localhost:19530"

# Create a MilvusClient instance
client = MilvusClient(uri=SERVER_ADDR)

# Define the collection schema
schema = client.create_schema(
    auto_id=False,
    enable_dynamic_fields=True,
)

# Add an INT64 field `age` that supports null values with default value 18
schema.add_field(field_name="age", datatype=DataType.INT64, nullable=True, default_value=18)
# Add a FLOAT field `price` that supports null values without default value
schema.add_field(field_name="price", datatype=DataType.FLOAT, nullable=True)
schema.add_field(field_name="pk", datatype=DataType.INT64, is_primary=True)
schema.add_field(field_name="embedding", datatype=DataType.FLOAT_VECTOR, dim=3)

import io.milvus.v2.client.ConnectConfig;
import io.milvus.v2.client.MilvusClientV2;

import io.milvus.v2.common.DataType;
import io.milvus.v2.service.collection.request.AddFieldReq;
import io.milvus.v2.service.collection.request.CreateCollectionReq;

MilvusClientV2 client = new MilvusClientV2(ConnectConfig.builder()
        .uri("http://localhost:19530")
        .build());
        
CreateCollectionReq.CollectionSchema schema = client.createSchema();
schema.setEnableDynamicField(true);

schema.addField(AddFieldReq.builder()
        .fieldName("age")
        .dataType(DataType.Int64)
        .isNullable(true)
        .defaultValue(18)
        .build());

schema.addField(AddFieldReq.builder()
        .fieldName("price")
        .dataType(DataType.Float)
        .isNullable(true)
        .build());

schema.addField(AddFieldReq.builder()
        .fieldName("pk")
        .dataType(DataType.Int64)
        .isPrimaryKey(true)
        .build());

schema.addField(AddFieldReq.builder()
        .fieldName("embedding")
        .dataType(DataType.FloatVector)
        .dimension(3)
        .build());

import { MilvusClient, DataType } from "@zilliz/milvus2-sdk-node";
const schema = [
  {
    name: "age",
    data_type: DataType.Int64,
  },
  {
    name: "price",
    data_type: DataType.Float,
  },
  {
    name: "pk",
    data_type: DataType.Int64,
    is_primary_key: true,
  },
  {
    name: "embedding",
    data_type: DataType.FloatVector,
    dim: 3,
  },
];

import (
    "context"
    "fmt"

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

schema := entity.NewSchema()
schema.WithField(entity.NewField().
    WithName("pk").
    WithDataType(entity.FieldTypeInt64).
    WithIsPrimaryKey(true),
).WithField(entity.NewField().
    WithName("embedding").
    WithDataType(entity.FieldTypeFloatVector).
    WithDim(3),
).WithField(entity.NewField().
    WithName("price").
    WithDataType(entity.FieldTypeFloat).
    WithNullable(true),
).WithField(entity.NewField().
    WithName("age").
    WithDataType(entity.FieldTypeInt64).
    WithNullable(true).
    WithDefaultValueLong(18),
)

export int64Field='{
    "fieldName": "age",
    "dataType": "Int64"
}'

export floatField='{
    "fieldName": "price",
    "dataType": "Float"
}'

export pkField='{
    "fieldName": "pk",
    "dataType": "Int64",
    "isPrimary": true
}'

export vectorField='{
    "fieldName": "embedding",
    "dataType": "FloatVector",
    "elementTypeParams": {
        "dim": 3
    }
}'

export schema="{
    \"autoID\": false,
    \"fields\": [
        $int64Field,
        $floatField,
        $pkField,
        $vectorField
    ]
}"

設定索引參數

索引有助於改善搜尋與查詢效能。在 Milvus 中，對向量欄位必須建立索引，但對標量值欄位則是可選的。

以下範例在向量欄位embedding 和標量欄位age 上建立索引，兩者都使用AUTOINDEX 索引類型。使用這種類型，Milvus 會根據資料類型自動選擇最適合的索引。您也可以自訂每個欄位的索引類型和參數。詳情請參閱Index Explained。

# Set index params

index_params = client.prepare_index_params()

# Index `age` with AUTOINDEX
index_params.add_index(
    field_name="age",
    index_type="AUTOINDEX",
    index_name="age_index"
)

# Index `embedding` with AUTOINDEX and specify similarity metric type
index_params.add_index(
    field_name="embedding",
    index_type="AUTOINDEX",  # Use automatic indexing to simplify complex index settings
    metric_type="COSINE"  # Specify similarity metric type, options include L2, COSINE, or IP
)

import io.milvus.v2.common.IndexParam;
import java.util.*;

List<IndexParam> indexes = new ArrayList<>();
indexes.add(IndexParam.builder()
        .fieldName("age")
        .indexType(IndexParam.IndexType.AUTOINDEX)
        .build());
        
indexes.add(IndexParam.builder()
        .fieldName("embedding")
        .indexType(IndexParam.IndexType.AUTOINDEX)
        .metricType(IndexParam.MetricType.COSINE)
        .build());

import { IndexType } from "@zilliz/milvus2-sdk-node";
const indexParams = [
  {
    field_name: "age",
    index_name: "inverted_index",
    index_type: IndexType.AUTOINDEX,
  },
  {
    field_name: "embedding",
    metric_type: "COSINE",
    index_type: IndexType.AUTOINDEX,
  },
];

indexOption1 := milvusclient.NewCreateIndexOption("my_collection", "embedding",
    index.NewAutoIndex(index.MetricType(entity.IP)))
indexOption2 := milvusclient.NewCreateIndexOption("my_collection", "age",
    index.NewInvertedIndex())

export indexParams='[
        {
            "fieldName": "age",
            "indexName": "inverted_index",
            "indexType": "AUTOINDEX"
        },
        {
            "fieldName": "embedding",
            "metricType": "COSINE",
            "indexType": "AUTOINDEX"
        }
    ]'

建立集合

定義模式和索引後，建立包含數字欄位的集合。

# Create Collection
client.create_collection(
    collection_name="my_collection",
    schema=schema,
    index_params=index_params
)

CreateCollectionReq requestCreate = CreateCollectionReq.builder()
        .collectionName("my_collection")
        .collectionSchema(schema)
        .indexParams(indexes)
        .build();
client.createCollection(requestCreate);

client.create_collection({
    collection_name: "my_collection",
    schema: schema,
    index_params: indexParams
})

err = client.CreateCollection(ctx,
    milvusclient.NewCreateCollectionOption("my_collection", schema).
        WithIndexOptions(indexOption1, indexOption2))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d "{
    \"collectionName\": \"my_collection\",
    \"schema\": $schema,
    \"indexParams\": $indexParams
}"

插入資料

建立集合後，插入符合模式的實體。

# Sample data
data = [
    {"age": 25, "price": 99.99, "pk": 1, "embedding": [0.1, 0.2, 0.3]},
    {"age": 30, "pk": 2, "embedding": [0.4, 0.5, 0.6]}, # `price` field is missing, which should be null
    {"age": None, "price": None, "pk": 3, "embedding": [0.2, 0.3, 0.1]},  # `age` should default to 18, `price` is null
    {"age": 45, "price": None, "pk": 4, "embedding": [0.9, 0.1, 0.4]},  # `price` is null
    {"age": None, "price": 59.99, "pk": 5, "embedding": [0.8, 0.5, 0.3]},  # `age` should default to 18
    {"age": 60, "price": None, "pk": 6, "embedding": [0.1, 0.6, 0.9]}  # `price` is null
]

client.insert(
    collection_name="my_collection",
    data=data
)

import com.google.gson.Gson;
import com.google.gson.JsonObject;

import io.milvus.v2.service.vector.request.InsertReq;
import io.milvus.v2.service.vector.response.InsertResp;

List<JsonObject> rows = new ArrayList<>();
Gson gson = new Gson();
rows.add(gson.fromJson("{\"age\": 25, \"price\": 99.99, \"pk\": 1, \"embedding\": [0.1, 0.2, 0.3]}", JsonObject.class));
rows.add(gson.fromJson("{\"age\": 30, \"pk\": 2, \"embedding\": [0.4, 0.5, 0.6]}", JsonObject.class));
rows.add(gson.fromJson("{\"age\": null, \"price\": null, \"pk\": 3, \"embedding\": [0.2, 0.3, 0.1]}", JsonObject.class));
rows.add(gson.fromJson("{\"age\": 45, \"price\": null, \"pk\": 4, \"embedding\": [0.9, 0.1, 0.4]}", JsonObject.class));
rows.add(gson.fromJson("{\"age\": null, \"price\": 59.99, \"pk\": 5, \"embedding\": [0.8, 0.5, 0.3]}", JsonObject.class));
rows.add(gson.fromJson("{\"age\": 60, \"price\": null, \"pk\": 6, \"embedding\": [0.1, 0.6, 0.9]}", JsonObject.class));

InsertResp insertR = client.insert(InsertReq.builder()
        .collectionName("my_collection")
        .data(rows)
        .build());

const data = [
  { age: 25, price: 99.99, pk: 1, embedding: [0.1, 0.2, 0.3] },
  { age: 30, price: 149.5, pk: 2, embedding: [0.4, 0.5, 0.6] },
  { age: 35, price: 199.99, pk: 3, embedding: [0.7, 0.8, 0.9] },
];

client.insert({
  collection_name: "my_collection",
  data: data,
});

column1, _ := column.NewNullableColumnFloat("price",
    []float32{99.99, 59.99},
    []bool{true, false, false, false, true, false})
column2, _ := column.NewNullableColumnInt64("age",
    []int64{25, 30, 45, 60},
    []bool{true, true, false, true, false, true})

_, err = client.Insert(ctx, milvusclient.NewColumnBasedInsertOption("my_collection").
    WithInt64Column("pk", []int64{1, 2, 3, 4, 5, 6}).
    WithFloatVectorColumn("embedding", 3, [][]float32{
        {0.1, 0.2, 0.3},
        {0.4, 0.5, 0.6},
        {0.2, 0.3, 0.1},
        {0.9, 0.1, 0.4},
        {0.8, 0.5, 0.3},
        {0.1, 0.6, 0.9},
    }).
    WithColumns(column1, column2),
)
if err != nil {
    fmt.Println(err.Error())
    // handle err
}

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/insert" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
    "data": [
        {"age": 25, "price": 99.99, "pk": 1, "embedding": [0.1, 0.2, 0.3]},
        {"age": 30, "price": 149.50, "pk": 2, "embedding": [0.4, 0.5, 0.6]},
        {"age": 35, "price": 199.99, "pk": 3, "embedding": [0.7, 0.8, 0.9]}       
    ],
    "collectionName": "my_collection"
}'

使用篩選表達式查詢

插入實體後，使用query 方法擷取符合指定篩選表達式的實體。

檢索age 大於 30 的實體：

filter = 'age > 30'

res = client.query(
    collection_name="my_collection",
    filter=filter,
    output_fields=["age", "price", "pk"]
)

print(res)

# Example output:
# data: [
#     "{'age': 45, 'price': None, 'pk': 4}",
#     "{'age': 60, 'price': None, 'pk': 6}"
# ]

import io.milvus.v2.service.vector.request.QueryReq;
import io.milvus.v2.service.vector.response.QueryResp;

String filter = "age > 30";

QueryResp resp = client.query(QueryReq.builder()
        .collectionName("my_collection")
        .filter(filter)
        .outputFields(Arrays.asList("age", "price", "pk"))
        .build());
System.out.println(resp.getQueryResults());

// Output
//
// [
//    QueryResp.QueryResult(entity={price=null, pk=4, age=45}), 
//    QueryResp.QueryResult(entity={price=null, pk=6, age=60})
// ]

client.query({
    collection_name: 'my_collection',
    filter: 'age > 30',
    output_fields: ['age', 'price', 'pk']
});

filter := "age > 30"
queryResult, err := client.Query(ctx, milvusclient.NewQueryOption("my_collection").
    WithFilter(filter).
    WithOutputFields("pk", "age", "price"))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}
fmt.Println("pk", queryResult.GetColumn("pk").FieldData().GetScalars())
fmt.Println("age", queryResult.GetColumn("age").FieldData().GetScalars())
fmt.Println("price", queryResult.GetColumn("price").FieldData().GetScalars())

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/query" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
    "collectionName": "my_collection",
    "filter": "age > 30",
    "outputFields": ["age","price", "pk"]
}'

## {"code":0,"cost":0,"data":[{"age":30,"pk":2,"price":149.5},{"age":35,"pk":3,"price":199.99}]}

擷取price 為 null 的實體：

filter = 'price is null'

res = client.query(
    collection_name="my_collection",
    filter=filter,
    output_fields=["age", "price", "pk"]
)

print(res)

# Example output:
# data: [
#     "{'age': 30, 'price': None, 'pk': 2}",
#     "{'age': 18, 'price': None, 'pk': 3}",
#     "{'age': 45, 'price': None, 'pk': 4}",
#     "{'age': 60, 'price': None, 'pk': 6}"
# ]

String filter = "price is null";

QueryResp resp = client.query(QueryReq.builder()
        .collectionName("my_collection")
        .filter(filter)
        .outputFields(Arrays.asList("age", "price", "pk"))
        .build());
System.out.println(resp.getQueryResults());

// Output
// [
//    QueryResp.QueryResult(entity={price=null, pk=2, age=30}), 
//    QueryResp.QueryResult(entity={price=null, pk=3, age=18}), 
//    QueryResp.QueryResult(entity={price=null, pk=4, age=45}), 
//    QueryResp.QueryResult(entity={price=null, pk=6, age=60})
// ]

// node
const filter = 'price is null';

const res = await client.query({
    collection_name:"my_collection",
    filter:filter,
    output_fields=["age", "price", "pk"]
});

console.log(res);

// Example output:
// data: [
//     "{'age': 18, 'price': None, 'pk': 3}",
//     "{'age': 18, 'price': 59.99, 'pk': 5}"
// ]

filter = "price is null"
queryResult, err = client.Query(ctx, milvusclient.NewQueryOption("my_collection").
    WithFilter(filter).
    WithOutputFields("pk", "age", "price"))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}
fmt.Println("pk", queryResult.GetColumn("pk"))
fmt.Println("age", queryResult.GetColumn("age"))
fmt.Println("price", queryResult.GetColumn("price"))

# restful
curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/query" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
  "collectionName": "my_collection",
  "filter": "price is null",
  "outputFields": ["age", "price", "pk"]
}'

要檢索age 的值為18 的實體，請使用下列表達式。由於age 的預設值是18 ，預期的結果應該包括age 明確設定為18 或age 設定為 null 的實體。

filter = 'age == 18'

res = client.query(
    collection_name="my_collection",
    filter=filter,
    output_fields=["age", "price", "pk"]
)

print(res)

# Example output:
# data: [
#     "{'age': 18, 'price': None, 'pk': 3}",
#     "{'age': 18, 'price': 59.99, 'pk': 5}"
# ]

String filter = "age == 18";

QueryResp resp = client.query(QueryReq.builder()
        .collectionName("my_collection")
        .filter(filter)
        .outputFields(Arrays.asList("age", "price", "pk"))
        .build());
System.out.println(resp.getQueryResults());

// Output
// [
//    QueryResp.QueryResult(entity={price=null, pk=3, age=18}), 
//    QueryResp.QueryResult(entity={price=59.99, pk=5, age=18})
// ]

// node
const filter = 'age == 18';

const res = await client.query({
    collection_name:"my_collection",
    filter:filter,
    output_fields=["age", "price", "pk"]
});

console.log(res);

// Example output:
// data: [
//     "{'age': 18, 'price': None, 'pk': 3}",
//     "{'age': 18, 'price': 59.99, 'pk': 5}"
// ]

filter = "age == 18"
queryResult, err = client.Query(ctx, milvusclient.NewQueryOption("my_collection").
    WithFilter(filter).
    WithOutputFields("pk", "age", "price"))
if err != nil {
    fmt.Println(err.Error())
    // handle error
}
fmt.Println("pk", queryResult.GetColumn("pk"))
fmt.Println("age", queryResult.GetColumn("age"))
fmt.Println("price", queryResult.GetColumn("price"))

# restful
curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/entities/query" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
--header "Request-Timeout: 10" \
-d '{
  "collectionName": "my_collection",
  "filter": "age == 18",
  "outputFields": ["age", "price", "pk"]
}'

使用篩選表達式的向量搜尋

除了基本的數字欄位篩選外，您也可以結合向量相似性搜尋與數字欄位篩選。例如，以下程式碼顯示如何將數字欄位篩選式加入向量搜尋：

filter = "25 <= age <= 35"

res = client.search(
    collection_name="my_collection",
    data=[[0.3, -0.6, 0.1]],
    limit=5,
    search_params={"params": {"nprobe": 10}},
    output_fields=["age","price"],
    filter=filter
)

print(res)

# Example output:
# data: [
#     "[{'id': 2, 'distance': -0.2016308456659317, 'entity': {'age': 30, 'price': None}}, {'id': 1, 'distance': -0.23643313348293304, 'entity': {'age': 25, 'price': 99.98999786376953}}]"
# ]

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;

String filter = "25 <= age <= 35";

SearchResp resp = client.search(SearchReq.builder()
        .collectionName("my_collection")
        .annsField("embedding")
        .data(Collections.singletonList(new FloatVec(new float[]{0.3f, -0.6f, 0.1f})))
        .topK(5)
        .outputFields(Arrays.asList("age", "price"))
        .filter(filter)
        .build());

System.out.println(resp.getSearchResults());

// Output
//
// [
//   [
//     SearchResp.SearchResult(entity={price=null, age=30}, score=-0.20163085, id=2),
//     SearchResp.SearchResult(entity={price=99.99, age=25}, score=-0.23643313, id=1)
//   ]
// ]

await client.search({
    collection_name: 'my_collection',
    data: [0.3, -0.6, 0.1],
    limit: 5,
    output_fields: ['age', 'price'],
    filter: '25 <= age <= 35'
});

queryVector := []float32{0.3, -0.6, 0.1}
filter = "25 <= age <= 35"

annParam := index.NewCustomAnnParam()
annParam.WithExtraParam("nprobe", 10)
resultSets, err := client.Search(ctx, milvusclient.NewSearchOption(
    "my_collection", // collectionName
    5,               // limit
    []entity.Vector{entity.FloatVector(queryVector)},
).WithANNSField("embedding").
    WithFilter(filter).
    WithAnnParam(annParam).
    WithOutputFields("age", "price"))
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("age: ", resultSet.GetColumn("age"))
    fmt.Println("price: ", resultSet.GetColumn("price"))
}

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": [
        [0.3, -0.6, 0.1]
    ],
    "annsField": "embedding",
    "limit": 5,
    "outputFields": ["age", "price"]
}'

## {"code":0,"cost":0,"data":[{"age":35,"distance":-0.19054288,"id":3,"price":199.99},{"age":30,"distance":-0.20163085,"id":2,"price":149.5},{"age":25,"distance":-0.2364331,"id":1,"price":99.99}]}

在這個範例中，我們首先定義一個查詢向量，並在搜尋過程中加入篩選條件25 <= age <= 35 。這可確保搜尋結果不僅與查詢向量相似，也符合指定的年齡範圍。如需詳細資訊，請參閱過濾。