Ban AutoIndex on BinaryVector (#21381)

Signed-off-by: Nzhenshan.cao <zhenshan.cao@zilliz.com>

internal/indexcoord/task.go

@@ -19,6 +19,9 @@ package indexcoord
 import (
 	"context"
 	"errors"
+	"fmt"
+
+	"github.com/milvus-io/milvus-proto/go-api/schemapb"
 
 	"go.uber.org/zap"
 
@@ -131,6 +134,11 @@ func (cit *CreateIndexTask) PreExecute(ctx context.Context) error {
 	if getIndexType(cit.req.GetIndexParams()) == diskAnnIndex && !cit.indexCoordClient.nodeManager.ClientSupportDisk() {
 		return errors.New("all IndexNodes do not support disk indexes, please verify")
 	}
+
+	if cit.req.GetFieldDataType() == schemapb.DataType_BinaryVector && Params.AutoIndexConfig.Enable {
+		return fmt.Errorf("can not create AutoIndex on binary vector")
+	}
+
 	return nil
 }
 

internal/proto/index_coord.proto

@@ -5,6 +5,7 @@ package milvus.proto.index;
 option go_package = "github.com/milvus-io/milvus/internal/proto/indexpb";
 
 import "common.proto";
+import "schema.proto";
 import "internal.proto";
 import "milvus.proto";
 
@@ -131,6 +132,7 @@ message CreateIndexRequest {
   uint64 timestamp = 6;
   bool  is_auto_index = 7;
   repeated common.KeyValuePair user_index_params = 8;
+  schema.DataType fieldDataType = 9;
 }
 
 message GetIndexInfoRequest {

internal/proto/internal.proto

@@ -117,17 +117,6 @@ message AlterAliasRequest{
   string alias = 4;
 }
 
-message CreateIndexRequest {
-  common.MsgBase base = 1;
-  string db_name = 2;
-  string collection_name = 3;
-  string field_name = 4;
-  int64 dbID = 5;
-  int64 collectionID = 6;
-  int64 fieldID = 7;
-  repeated common.KeyValuePair extra_params = 8;
-}
-
 enum InsertDataVersion {
   // 0 must refer to row-based format, since it's the first version in Milvus.
   RowBased = 0;

internal/proxy/task.go

@@ -202,6 +202,9 @@ func (cct *createCollectionTask) PreExecute(ctx context.Context) error {
 			if err != nil {
 				return err
 			}
+			if Params.AutoIndexConfig.Enable && field.DataType == schemapb.DataType_BinaryVector {
+				return fmt.Errorf("can not speficy binary vector when enabled with AutoIndex")
+			}
 		}
 		// valid max length per row parameters
 		// if max_length not specified, return error

internal/proxy/task_index.go

@@ -282,6 +282,11 @@ func (cit *createIndexTask) PreExecute(ctx context.Context) error {
 	if err != nil {
 		return err
 	}
+
+	if field.GetDataType() == schemapb.DataType_BinaryVector && Params.AutoIndexConfig.Enable {
+		return fmt.Errorf("can not create AutoIndex on binary vector")
+	}
+
 	cit.fieldSchema = field
 	// check index param, not accurate, only some static rules
 	err = cit.parseIndexParams()
@@ -318,6 +323,7 @@ func (cit *createIndexTask) Execute(ctx context.Context) error {
 		IsAutoIndex:     cit.isAutoIndex,
 		UserIndexParams: cit.req.GetExtraParams(),
 		Timestamp:       cit.BeginTs(),
+		FieldDataType:   cit.fieldSchema.GetDataType(),
 	}
 	cit.result, err = cit.indexCoord.CreateIndex(ctx, req)
 	if err != nil {

internal/rootcoord/create_collection_task.go

@@ -54,7 +54,6 @@ func (t *createCollectionTask) validate() error {
 	if t.Req.GetShardsNum() >= maxShardNum {
 		return fmt.Errorf("shard num (%d) exceeds limit (%d)", t.Req.GetShardsNum(), maxShardNum)
 	}
-
 	return nil
 }
 
@@ -67,6 +66,15 @@ func hasSystemFields(schema *schemapb.CollectionSchema, systemFields []string) b
 	return false
 }
 
+func hasBinaryVecField(schema *schemapb.CollectionSchema) bool {
+	for _, field := range schema.GetFields() {
+		if field.GetDataType() == schemapb.DataType_BinaryVector {
+			return true
+		}
+	}
+	return false
+}
+
 func (t *createCollectionTask) validateSchema(schema *schemapb.CollectionSchema) error {
 	if t.Req.GetCollectionName() != schema.GetName() {
 		return fmt.Errorf("collection name = %s, schema.Name=%s", t.Req.GetCollectionName(), schema.Name)
@@ -74,6 +82,10 @@ func (t *createCollectionTask) validateSchema(schema *schemapb.CollectionSchema)
 	if hasSystemFields(schema, []string{RowIDFieldName, TimeStampFieldName}) {
 		return fmt.Errorf("schema contains system field: %s, %s", RowIDFieldName, TimeStampFieldName)
 	}
+
+	if Params.AutoIndexConfig.Enable && hasBinaryVecField(schema) {
+		return fmt.Errorf("can not speficy binary vector when enabled with AutoIndex")
+	}
 	return nil
 }