mirror of
https://gitee.com/milvus-io/milvus.git
synced 2025-12-28 14:35:27 +08:00
related: #45993 This commit extends nullable vector support to the proxy layer, querynode, and adds comprehensive validation, search reduce, and field data handling for nullable vectors with sparse storage. Proxy layer changes: - Update validate_util.go checkAligned() with getExpectedVectorRows() helper to validate nullable vector field alignment using valid data count - Update checkFloatVectorFieldData/checkSparseFloatVectorFieldData for nullable vector validation with proper row count expectations - Add FieldDataIdxComputer in typeutil/schema.go for logical-to-physical index translation during search reduce operations - Update search_reduce_util.go reduceSearchResultData to use idxComputers for correct field data indexing with nullable vectors - Update task.go, task_query.go, task_upsert.go for nullable vector handling - Update msg_pack.go with nullable vector field data processing QueryNode layer changes: - Update segments/result.go for nullable vector result handling - Update segments/search_reduce.go with nullable vector offset translation Storage and index changes: - Update data_codec.go and utils.go for nullable vector serialization - Update indexcgowrapper/dataset.go and index.go for nullable vector indexing Utility changes: - Add FieldDataIdxComputer struct with Compute() method for efficient logical-to-physical index mapping across multiple field data - Update EstimateEntitySize() and AppendFieldData() with fieldIdxs parameter - Update funcutil.go with nullable vector support functions <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Full support for nullable vector fields (float, binary, float16, bfloat16, int8, sparse) across ingest, storage, indexing, search and retrieval; logical↔physical offset mapping preserves row semantics. * Client: compaction control and compaction-state APIs. * **Bug Fixes** * Improved validation for adding vector fields (nullable + dimension checks) and corrected search/query behavior for nullable vectors. * **Chores** * Persisted validity maps with indexes and on-disk formats. * **Tests** * Extensive new and updated end-to-end nullable-vector tests. <sub>✏️ Tip: You can customize this high-level summary in your review settings.</sub> <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: marcelo-cjl <marcelo.chen@zilliz.com>
336 lines
8.4 KiB
Go
336 lines
8.4 KiB
Go
// Licensed to the LF AI & Data foundation under one
|
|
// or more contributor license agreements. See the NOTICE file
|
|
// distributed with this work for additional information
|
|
// regarding copyright ownership. The ASF licenses this file
|
|
// to you under the Apache License, Version 2.0 (the
|
|
// "License"); you may not use this file except in compliance
|
|
// with the License. You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
|
|
package column
|
|
|
|
import (
|
|
"github.com/cockroachdb/errors"
|
|
"github.com/samber/lo"
|
|
|
|
"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
|
|
"github.com/milvus-io/milvus/client/v2/entity"
|
|
)
|
|
|
|
type GColumn[T any] interface {
|
|
Value(idx int) T
|
|
AppendValue(v T)
|
|
}
|
|
|
|
var _ Column = (*genericColumnBase[any])(nil)
|
|
|
|
// genericColumnBase implements `Column` interface
|
|
// it provides the basic function for each scalar params
|
|
type genericColumnBase[T any] struct {
|
|
name string
|
|
fieldType entity.FieldType
|
|
values []T
|
|
|
|
// nullable related fields
|
|
// note that nullable must be set to true explicitly
|
|
nullable bool
|
|
validData []bool
|
|
// nullable column could be presented in two modes
|
|
// - compactMode, in which all valid data are compacted into one slice
|
|
// - sparseMode, in which valid data are located in its index position
|
|
// while invalid one are filled with zero value.
|
|
// for Milvus 2.5.x and before, insert request shall be in compactMode while
|
|
// search & query results are formed in sparseMode
|
|
// this flag indicates which form current column are in and peform validation
|
|
// or conversion logical based on it
|
|
sparseMode bool
|
|
// indexMapping stores the compact-sparse mapping
|
|
indexMapping []int
|
|
}
|
|
|
|
// Name returns column name.
|
|
func (c *genericColumnBase[T]) Name() string {
|
|
return c.name
|
|
}
|
|
|
|
// Type returns corresponding field type.
|
|
// note that: it is not necessary to be 1-on-1 mapping
|
|
// say, `[]byte` could be lots of field type.
|
|
func (c *genericColumnBase[T]) Type() entity.FieldType {
|
|
return c.fieldType
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Len() int {
|
|
if c.validData != nil {
|
|
return len(c.validData)
|
|
}
|
|
return len(c.values)
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) AppendValue(a any) error {
|
|
if a == nil {
|
|
return c.AppendNull()
|
|
}
|
|
v, ok := a.(T)
|
|
if !ok {
|
|
return errors.Newf("unexpected append value type %T, field type %v", a, c.fieldType)
|
|
}
|
|
c.values = append(c.values, v)
|
|
if c.nullable {
|
|
c.validData = append(c.validData, true)
|
|
c.indexMapping = append(c.indexMapping, len(c.values)-1)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Slice(start, end int) Column {
|
|
return c.slice(start, end)
|
|
}
|
|
|
|
// WARNING: this methods works only for sparse mode column
|
|
func (c *genericColumnBase[T]) slice(start, end int) *genericColumnBase[T] {
|
|
l := c.Len()
|
|
if start > l {
|
|
start = l
|
|
}
|
|
if end == -1 || end > l {
|
|
end = l
|
|
}
|
|
result := &genericColumnBase[T]{
|
|
name: c.name,
|
|
fieldType: c.fieldType,
|
|
values: c.values[start:end],
|
|
nullable: c.nullable,
|
|
sparseMode: c.sparseMode,
|
|
}
|
|
if c.nullable {
|
|
result.validData = c.validData[start:end]
|
|
}
|
|
return result
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) FieldData() *schemapb.FieldData {
|
|
fd := values2FieldData(c.values, c.fieldType, 0)
|
|
fd.FieldName = c.name
|
|
fd.Type = schemapb.DataType(c.fieldType)
|
|
if c.nullable {
|
|
fd.ValidData = c.validData
|
|
}
|
|
return fd
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) rangeCheck(idx int) error {
|
|
if idx < 0 || idx >= c.Len() {
|
|
return errors.Newf("index %d out of range[0, %d)", idx, c.Len())
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Get(idx int) (any, error) {
|
|
idx = c.valueIndex(idx)
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return nil, err
|
|
}
|
|
return c.values[idx], nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) GetAsInt64(idx int) (int64, error) {
|
|
idx = c.valueIndex(idx)
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return 0, err
|
|
}
|
|
return value2Type[T, int64](c.values[idx])
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) GetAsString(idx int) (string, error) {
|
|
idx = c.valueIndex(idx)
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return "", err
|
|
}
|
|
return value2Type[T, string](c.values[idx])
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) GetAsDouble(idx int) (float64, error) {
|
|
idx = c.valueIndex(idx)
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return 0, err
|
|
}
|
|
return value2Type[T, float64](c.values[idx])
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) GetAsBool(idx int) (bool, error) {
|
|
idx = c.valueIndex(idx)
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return false, err
|
|
}
|
|
return value2Type[T, bool](c.values[idx])
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Value(idx int) (T, error) {
|
|
idx = c.valueIndex(idx)
|
|
var z T
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return z, err
|
|
}
|
|
return c.values[idx], nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) valueIndex(idx int) int {
|
|
if !c.nullable || c.sparseMode {
|
|
return idx
|
|
}
|
|
return c.indexMapping[idx]
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Data() []T {
|
|
return c.values
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) MustValue(idx int) T {
|
|
idx = c.valueIndex(idx)
|
|
if idx < 0 || idx > c.Len() {
|
|
panic("index out of range")
|
|
}
|
|
return c.values[idx]
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) AppendNull() error {
|
|
if !c.nullable {
|
|
return errors.New("append null to not nullable column")
|
|
}
|
|
|
|
c.validData = append(c.validData, false)
|
|
if !c.sparseMode {
|
|
c.indexMapping = append(c.indexMapping, -1)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) IsNull(idx int) (bool, error) {
|
|
if err := c.rangeCheck(idx); err != nil {
|
|
return false, err
|
|
}
|
|
if !c.nullable {
|
|
return false, nil
|
|
}
|
|
return !c.validData[idx], nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) Nullable() bool {
|
|
return c.nullable
|
|
}
|
|
|
|
// SetNullable update the nullable flag and change the valid data array according to the flag value.
|
|
// NOTE: set nullable to false will erase all the validData previously set.
|
|
func (c *genericColumnBase[T]) SetNullable(nullable bool) {
|
|
c.nullable = nullable
|
|
// initialize validData only when
|
|
if c.nullable && c.validData == nil {
|
|
// set valid flag for all exisiting values
|
|
c.validData = lo.RepeatBy(len(c.values), func(_ int) bool { return true })
|
|
}
|
|
|
|
if !c.nullable {
|
|
c.validData = nil
|
|
}
|
|
}
|
|
|
|
// ValidateNullable performs the sanity check for nullable column.
|
|
// it checks the length of data and the valid number indicated by validData slice,
|
|
// which shall be the same by definition
|
|
func (c *genericColumnBase[T]) ValidateNullable() error {
|
|
// skip check if column not nullable
|
|
if !c.nullable {
|
|
return nil
|
|
}
|
|
|
|
if c.sparseMode {
|
|
return c.validateNullableSparse()
|
|
}
|
|
return c.validateNullableCompact()
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) validateNullableCompact() error {
|
|
// count valid entries
|
|
var validCnt int
|
|
c.indexMapping = make([]int, len(c.validData))
|
|
for idx, v := range c.validData {
|
|
if v {
|
|
c.indexMapping[idx] = validCnt
|
|
validCnt++
|
|
} else {
|
|
c.indexMapping[idx] = -1
|
|
}
|
|
}
|
|
if validCnt != len(c.values) {
|
|
return errors.Newf("values number(%d) does not match valid count(%d)", len(c.values), validCnt)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) validateNullableSparse() error {
|
|
if len(c.validData) != len(c.values) {
|
|
return errors.Newf("values number (%d) does not match valid data len(%d)", len(c.values), len(c.validData))
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) CompactNullableValues() {
|
|
if !c.nullable || !c.sparseMode {
|
|
return
|
|
}
|
|
|
|
c.indexMapping = make([]int, len(c.validData))
|
|
var cnt int
|
|
for idx, valid := range c.validData {
|
|
if !valid {
|
|
c.indexMapping[idx] = -1
|
|
continue
|
|
}
|
|
c.values[cnt] = c.values[idx]
|
|
c.indexMapping[idx] = cnt
|
|
cnt++
|
|
}
|
|
c.values = c.values[0:cnt]
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) ValidCount() int {
|
|
if !c.nullable || len(c.validData) == 0 {
|
|
return len(c.values)
|
|
}
|
|
count := 0
|
|
for _, v := range c.validData {
|
|
if v {
|
|
count++
|
|
}
|
|
}
|
|
return count
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) withValidData(validData []bool) {
|
|
if len(validData) > 0 {
|
|
c.nullable = true
|
|
c.validData = validData
|
|
}
|
|
}
|
|
|
|
func (c *genericColumnBase[T]) base() *genericColumnBase[T] {
|
|
return c
|
|
}
|
|
|
|
type ColumnOption[T any] func(*genericColumnBase[T])
|
|
|
|
// WithSparseNullableMode returns a ColumnOption that sets the sparse mode for the column.
|
|
func WithSparseNullableMode[T any](flag bool) ColumnOption[T] {
|
|
return func(c *genericColumnBase[T]) {
|
|
c.sparseMode = flag
|
|
}
|
|
}
|