Create Collection

You can create a collection by defining its schema, index parameters, metric type, and whether to load it upon creation. This page introduces how to create a collection from scratch.

Overview

A collection is a two-dimensional table with fixed columns and variant rows. Each column represents a field, and each row represents an entity. A schema is required to implement such structural data management. Every entity to insert has to meet the constraints defined in the schema.

You can determine every aspect of a collection, including its schema, index parameters, metric type, and whether to load it upon creation to ensure that the collection fully meets your requirements.

To create a collection, you need to

• Create schema

• Set index parameters (Optional)

• Create collection

Create Schema

A schema defines the data structure of a collection. When creating a collection, you need to design the schema based on your requirements. For details, refer to Schema Explained.

The following code snippets create a schema with the enabled dynamic field and three mandatory fields named my_id, my_vector, and my_varchar.

# 3. Create a collection in customized setup mode
from pymilvus import MilvusClient, DataType

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

# 3.1. Create schema
schema = MilvusClient.create_schema(
    auto_id=False,
    enable_dynamic_field=True,
)

# 3.2. Add fields to schema
schema.add_field(field_name="my_id", datatype=DataType.INT64, is_primary=True)
schema.add_field(field_name="my_vector", datatype=DataType.FLOAT_VECTOR, dim=5)
schema.add_field(field_name="my_varchar", datatype=DataType.VARCHAR, max_length=512)

import io.milvus.v2.common.DataType;
import io.milvus.v2.client.ConnectConfig;
import io.milvus.v2.client.MilvusClientV2;
import io.milvus.v2.service.collection.request.AddFieldReq;
import io.milvus.v2.service.collection.request.CreateCollectionReq;

String CLUSTER_ENDPOINT = "http://localhost:19530";
String TOKEN = "root:Milvus";

// 1. Connect to Milvus server
ConnectConfig connectConfig = ConnectConfig.builder()
        .uri(CLUSTER_ENDPOINT)
        .token(TOKEN)
        .build();

MilvusClientV2 client = new MilvusClientV2(connectConfig);

// 3. Create a collection in customized setup mode

// 3.1 Create schema
CreateCollectionReq.CollectionSchema schema = client.createSchema();

// 3.2 Add fields to schema
schema.addField(AddFieldReq.builder()
        .fieldName("my_id")
        .dataType(DataType.Int64)
        .isPrimaryKey(true)
        .autoID(false)
        .build());

schema.addField(AddFieldReq.builder()
        .fieldName("my_vector")
        .dataType(DataType.FloatVector)
        .dimension(5)
        .build());

schema.addField(AddFieldReq.builder()
        .fieldName("my_varchar")
        .dataType(DataType.VarChar)
        .maxLength(512)
        .build());

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

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

// 3. Create a collection in customized setup mode
// 3.1 Define fields
const fields = [
    {
        name: "my_id",
        data_type: DataType.Int64,
        is_primary_key: true,
        auto_id: false
    },
    {
        name: "my_vector",
        data_type: DataType.FloatVector,
        dim: 5
    },
    {
        name: "my_varchar",
        data_type: DataType.VarChar,
        max_length: 512
    }
]

import "github.com/milvus-io/milvus/client/v2/entity"

schema := entity.NewSchema().WithDynamicFieldEnabled(true).
        WithField(entity.NewField().WithName("my_id").WithIsAutoID(true).WithDataType(entity.FieldTypeInt64).WithIsPrimaryKey(true)).
        WithField(entity.NewField().WithName("my_vector").WithDataType(entity.FieldTypeFloatVector).WithDim(5)).
        WithField(entity.NewField().WithName("my_varchar").WithDataType(entity.FieldTypeVarChar).WithMaxLength(512))thDim(5))

export schema='{
        "autoId": false,
        "enabledDynamicField": false,
        "fields": [
            {
                "fieldName": "my_id",
                "dataType": "Int64",
                "isPrimary": true
            },
            {
                "fieldName": "my_vector",
                "dataType": "FloatVector",
                "elementTypeParams": {
                    "dim": "5"
                }
            },
            {
                "fieldName": "my_varchar",
                "dataType": "VarChar",
                "elementTypeParams": {
                    "max_length": 512
                }
            }
        ]
    }'

(Optional) Set Index Parameters

Creating an index on a specific field accelerates the search against this field. An index records the order of entities within a collection. As shown in the following code snippets, you can use metric_type and index_type to select appropriate ways for Milvus to index a field and measure similarities between vector embeddings.

In Milvus, you can use AUTOINDEX as the index type for all vector fields, and one of COSINE, L2, and IP as the metric type based on your needs.

As demonstrated in the above code snippet, you need to set both the index type and metric type for vector fields and only the index type for the scalar fields. Indexes are mandatory for vector fields, and you are advised to create indexes on scalar fields frequently used in filtering conditions.

For details, refer to Indexes.

# 3.3. Prepare index parameters
index_params = client.prepare_index_params()

# 3.4. Add indexes
index_params.add_index(
    field_name="my_id",
    index_type="STL_SORT"
)

index_params.add_index(
    field_name="my_vector", 
    index_type="AUTOINDEX",
    metric_type="COSINE"
)

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

// 3.3 Prepare index parameters
IndexParam indexParamForIdField = IndexParam.builder()
        .fieldName("my_id")
        .indexType(IndexParam.IndexType.STL_SORT)
        .build();

IndexParam indexParamForVectorField = IndexParam.builder()
        .fieldName("my_vector")
        .indexType(IndexParam.IndexType.AUTOINDEX)
        .metricType(IndexParam.MetricType.COSINE)
        .build();

List<IndexParam> indexParams = new ArrayList<>();
indexParams.add(indexParamForIdField);
indexParams.add(indexParamForVectorField);

// 3.2 Prepare index parameters
const index_params = [{
    field_name: "my_id",
    index_type: "STL_SORT"
},{
    field_name: "my_vector",
    index_type: "AUTOINDEX",
    metric_type: "COSINE"
}]

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

indexOptions := []client.CreateIndexOption{
    client.NewCreateIndexOption(collectionName, "my_vector", index.NewAutoIndex(entity.COSINE)).WithIndexName("my_vector"),
    client.NewCreateIndexOption(collectionName, "my_id", index.NewSortedIndex()).WithIndexName("my_id"),
}

export indexParams='[
        {
            "fieldName": "my_vector",
            "metricType": "COSINE",
            "indexName": "my_vector",
            "indexType": "AUTOINDEX"
        },
        {
            "fieldName": "my_id",
            "indexName": "my_id",
            "indexType": "STL_SORT"
        }
    ]'

Create Collection

If you have created a collection with index parameters, Milvus automatically loads the collection upon its creation. In this case, all fields mentioned in the index parameters are indexed.

The following code snippets demonstrate how to create the collection with index parameters and check its load status.

# 3.5. Create a collection with the index loaded simultaneously
client.create_collection(
    collection_name="customized_setup_1",
    schema=schema,
    index_params=index_params
)

res = client.get_load_state(
    collection_name="customized_setup_1"
)

print(res)

# Output
#
# {
#     "state": "<LoadState: Loaded>"
# }

import io.milvus.v2.service.collection.request.CreateCollectionReq;
import io.milvus.v2.service.collection.request.GetLoadStateReq;

// 3.4 Create a collection with schema and index parameters
CreateCollectionReq customizedSetupReq1 = CreateCollectionReq.builder()
        .collectionName("customized_setup_1")
        .collectionSchema(schema)
        .indexParams(indexParams)
        .build();

client.createCollection(customizedSetupReq1);

// 3.5 Get load state of the collection
GetLoadStateReq customSetupLoadStateReq1 = GetLoadStateReq.builder()
        .collectionName("customized_setup_1")
        .build();

Boolean loaded = client.getLoadState(customSetupLoadStateReq1);
System.out.println(loaded);

// Output:
// true

// 3.3 Create a collection with fields and index parameters
res = await client.createCollection({
    collection_name: "customized_setup_1",
    fields: fields,
    index_params: index_params,
})

console.log(res.error_code)  

// Output
// 
// Success
// 

res = await client.getLoadState({
    collection_name: "customized_setup_1"
})

console.log(res.state)

// Output
// 
// LoadStateLoaded
// 

import "github.com/milvus-io/milvus/client/v2"

err := cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_1", schema).
    WithIndexOptions(indexOptions...),
)
if err != nil {
    // handle error
}
fmt.Println("collection created")

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

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_1\",
    \"schema\": $schema,
    \"indexParams\": $indexParams
}"

You can also create a collection without any index parameters and add them afterward. In this case, Milvus does not load the collection upon its creation. For details on how to create indexes for an existing collection, refer to Index Explained.

The following code snippet demonstrates how to create a collection without an index, and the load status of the collection remains unloaded upon creation.

# 3.6. Create a collection and index it separately
client.create_collection(
    collection_name="customized_setup_2",
    schema=schema,
)

res = client.get_load_state(
    collection_name="customized_setup_2"
)

print(res)

# Output
#
# {
#     "state": "<LoadState: NotLoad>"
# }

// 3.6 Create a collection and index it separately
CreateCollectionReq customizedSetupReq2 = CreateCollectionReq.builder()
    .collectionName("customized_setup_2")
    .collectionSchema(schema)
    .build();

client.createCollection(customizedSetupReq2);

GetLoadStateReq customSetupLoadStateReq2 = GetLoadStateReq.builder()
        .collectionName("customized_setup_2")
        .build();
        
Boolean loaded = client.getLoadState(customSetupLoadStateReq2);
System.out.println(loaded);

// Output:
// false

// 3.4 Create a collection and index it seperately
res = await client.createCollection({
    collection_name: "customized_setup_2",
    fields: fields,
})

console.log(res.error_code)

// Output
// 
// Success
// 

res = await client.getLoadState({
    collection_name: "customized_setup_2"
})

console.log(res.state)

// Output
// 
// LoadStateNotLoad
// 

import "github.com/milvus-io/milvus/client/v2"

err := cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_2", schema))
if err != nil {
    // handle error
}
fmt.Println("collection created")

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

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_2\",
    \"schema\": $schema
}"

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/get_load_state" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_2\"
}"

Milvus also provides a way for you to create a collection instantly. For details, refer to Create Collection Instantly.

Set Collection Properties

You can set properties for the collection to create to make it fit into your service. The applicable properties are as follows.

Set Shard Number

Shards are horizontal slices of a collection. Each shard corresponds to a data input channel. Every collection has a shard by default. You can set the appropriate number of shards when creating a collection based on the expected throughput and the volume of the data to insert into the collection.

In common cases, consider increasing the shard number by one every time the expected throughput increases by 500 MB/s or the volume of data to insert increases by 100 GB. This suggestion is based on our own experience and may not completely fit in your application scenarios. You can tune this number to fit your own needs or just use the default value.

The following code snippet demonstrates how to set the shard number when you create a collection.

# With shard number
client.create_collection(
    collection_name="customized_setup_3",
    schema=schema,
    # highlight-next-line
    num_shards=1
)

// With shard number
CreateCollectionReq customizedSetupReq3 = CreateCollectionReq.builder()
    .collectionName("customized_setup_3")
    .collectionSchema(collectionSchema)
    // highlight-next-line
    .numShards(1)
    .build();
client.createCollection(customizedSetupReq3);

const createCollectionReq = {
    collection_name: "customized_setup_3",
    schema: schema,
    // highlight-next-line
    shards_num: 1
}

import "github.com/milvus-io/milvus/client/v2"

err := cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_3", schema).WithShardNum(1))
if err != nil {
    // handle error
}
fmt.Println("collection created")

export params='{
    "shardsNum": 1
}'

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

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_3\",
    \"schema\": $schema,
    \"params\": $params
}"

Enable mmap

Milvus enables mmap on all collections by default, allowing Milvus to map raw field data into memory instead of fully loading them. This reduces memory footprints and increases collection capacity. For details on mmap, refer to Use mmap.

# With mmap
client.create_collection(
    collection_name="customized_setup_4",
    schema=schema,
    # highlight-next-line
    enable_mmap=False
)

import io.milvus.param.Constant;

// With MMap
CreateCollectionReq customizedSetupReq4 = CreateCollectionReq.builder()
        .collectionName("customized_setup_4")
        .collectionSchema(schema)
        // highlight-next-line
        .property(Constant.MMAP_ENABLED, "false")
        .build();
client.createCollection(customizedSetupReq4);

client.create_collection({
    collection_name: "customized_setup_4",
    schema: schema,
     properties: {
        'mmap.enabled': true,
     },
})

import (
    "github.com/milvus-io/milvus/client/v2"
    "github.com/milvus-io/milvus/pkg/common"
)

err := cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_4", schema).WithProperty(common.MmapEnabledKey, true))
if err != nil {
    // handle error
}
fmt.Println("collection created")

# Currently not available for REST

Set Collection TTL

If a collection needs to be dropped for a specific period, consider setting its Time-To-Live (TTL) in seconds. Once the TTL times out, Milvus deletes entities in the collection and drops the collection. The deletion is asynchronous, indicating that searches and queries are still possible before the deletion is complete.

The following code snippet sets the TTL to one day (86400 seconds). You are advised to set the TTL to a couple of days at minimum.

# With TTL
client.create_collection(
    collection_name="customized_setup_5",
    schema=schema,
    # highlight-start
    properties={
        "collection.ttl.seconds": 86400
    }
    # highlight-end
)

import io.milvus.param.Constant;

// With TTL
CreateCollectionReq customizedSetupReq5 = CreateCollectionReq.builder()
        .collectionName("customized_setup_5")
        .collectionSchema(schema)
        // highlight-next-line
        .property(Constant.TTL_SECONDS, "86400")
        .build();
client.createCollection(customizedSetupReq5);

const createCollectionReq = {
    collection_name: "customized_setup_5",
    schema: schema,
    // highlight-start
    properties: {
        "collection.ttl.seconds": 86400
    }
    // highlight-end
}

import (
    "github.com/milvus-io/milvus/client/v2"
    "github.com/milvus-io/milvus/pkg/common"
)

err = cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_5", schema).
        WithProperty(common.CollectionTTLConfigKey, 86400)) //  TTL in seconds
if err != nil {
        // handle error
}
fmt.Println("collection created")

export params='{
    "ttlSeconds": 86400
}'

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

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_5\",
    \"schema\": $schema,
    \"params\": $params
}"

Set Consistency Level

When creating a collection, you can set the consistency level for searches and queries in the collection. You can also change the consistency level of the collection during a specific search or query.

# With consistency level
client.create_collection(
    collection_name="customized_setup_6",
    schema=schema,
    # highlight-next
    consistency_level="Bounded",
)

import io.milvus.v2.common.ConsistencyLevel;

// With consistency level
CreateCollectionReq customizedSetupReq6 = CreateCollectionReq.builder()
        .collectionName("customized_setup_6")
        .collectionSchema(schema)
        // highlight-next-line
        .consistencyLevel(ConsistencyLevel.BOUNDED)
        .build();
client.createCollection(customizedSetupReq6);

const createCollectionReq = {
    collection_name: "customized_setup_6",
    schema: schema,
    // highlight-next
    consistency_level: "Bounded",
    // highlight-end
}

client.createCollection(createCollectionReq);

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

err := cli.CreateCollection(ctx, client.NewCreateCollectionOption("customized_setup_6", schema).
    WithConsistencyLevel(entity.ClBounded))
if err != nil {
    // handle error
}
fmt.Println("collection created")

export params='{
    "consistencyLevel": "Bounded"
}'

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

curl --request POST \
--url "${CLUSTER_ENDPOINT}/v2/vectordb/collections/create" \
--header "Authorization: Bearer ${TOKEN}" \
--header "Content-Type: application/json" \
-d "{
    \"collectionName\": \"customized_setup_6\",
    \"schema\": $schema,
    \"params\": $params
}"

For more on consistency levels, see Consistency Level.

Enable Dynamic Field

The dynamic field in a collection is a reserved JavaScript Object Notation (JSON) field named $meta. Once you have enabled this field, Milvus saves all non-schema-defined fields carried in each entity and their values as key-value pairs in the reserved field.

For details on how to use the dynamic field, refer to Dynamic Field.