lsm/seaweedfs
1
0
Fork 0
mirror of https://github.com/seaweedfs/seaweedfs.git synced 2025-09-21 09:28:01 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

use schema instead of inferred result types

Browse Source
This commit is contained in:
chrislu
2025-09-02 20:59:13 -07:00
parent 316d1cdda7
commit e14a316aeb
5 changed files with 105 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
weed/command/sql.go
View File

@@ -299,7 +299,6 @@ func runInteractiveShell(ctx *SQLContext) bool {
queryBuffer.Reset()
continue
}
}
// Handle output format switching
if strings.HasPrefix(strings.ToUpper(cleanQuery), "\\FORMAT ") {

4
weed/query/engine/hybrid_message_scanner.go
View File

@@ -782,6 +782,8 @@ func (hms *HybridMessageScanner) ConvertToSQLResult(results []HybridScanResult,
return &QueryResult{
Columns: columns,
Rows: [][]sqltypes.Value{},
Database: hms.topic.Namespace,
Table: hms.topic.Name,
}
}
@@ -826,6 +828,8 @@ func (hms *HybridMessageScanner) ConvertToSQLResult(results []HybridScanResult,
return &QueryResult{
Columns: columns,
Rows: rows,
Database: hms.topic.Namespace,
Table: hms.topic.Name,
}
}

3
weed/query/engine/types.go
View File

@@ -31,4 +31,7 @@ type QueryResult struct {
Rows [][]sqltypes.Value `json:"rows"`
Error error `json:"error,omitempty"`
ExecutionPlan *QueryExecutionPlan `json:"execution_plan,omitempty"`
// Schema information for type inference (optional)
Database string `json:"database,omitempty"`
Table string `json:"table,omitempty"`
}

100
weed/server/postgres/protocol.go
View File

@@ -9,6 +9,7 @@ import (
"strings"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/pb/schema_pb"
"github.com/seaweedfs/seaweedfs/weed/query/engine"
"github.com/seaweedfs/seaweedfs/weed/query/sqltypes"
"github.com/seaweedfs/seaweedfs/weed/util/version"
@@ -202,7 +203,7 @@ func (s *PostgreSQLServer) handleSimpleQuery(session *PostgreSQLSession, query s
// Send results for this statement
if len(result.Columns) > 0 {
// Send row description
err = s.sendRowDescription(session, result.Columns, result.Rows)
err = s.sendRowDescription(session, result)
if err != nil {
return err
}
@@ -324,8 +325,12 @@ func (s *PostgreSQLServer) sendSystemQueryResult(session *PostgreSQLSession, res
sqlRows = append(sqlRows, sqlRow)
}
// Send row description
err := s.sendRowDescription(session, columns, sqlRows)
// Send row description (create a temporary QueryResult for consistency)
tempResult := &engine.QueryResult{
Columns: columns,
Rows: sqlRows,
}
err := s.sendRowDescription(session, tempResult)
if err != nil {
return err
}
@@ -418,7 +423,11 @@ func (s *PostgreSQLServer) handleDescribe(session *PostgreSQLSession, msgBody []
glog.V(2).Infof("PostgreSQL Describe %c (ID: %d): %s", objectType, session.processID, objectName)
// For now, send empty row description
return s.sendRowDescription(session, []string{}, [][]sqltypes.Value{})
tempResult := &engine.QueryResult{
Columns: []string{},
Rows: [][]sqltypes.Value{},
}
return s.sendRowDescription(session, tempResult)
}
// handleClose processes a Close message
@@ -509,13 +518,13 @@ func (s *PostgreSQLServer) sendReadyForQuery(session *PostgreSQLSession) error {
}
// sendRowDescription sends row description message
func (s *PostgreSQLServer) sendRowDescription(session *PostgreSQLSession, columns []string, rows [][]sqltypes.Value) error {
func (s *PostgreSQLServer) sendRowDescription(session *PostgreSQLSession, result *engine.QueryResult) error {
msg := make([]byte, 0)
msg = append(msg, PG_RESP_ROW_DESC)
// Calculate message length
length := 4 + 2 // length + field count
for _, col := range columns {
for _, col := range result.Columns {
length += len(col) + 1 + 4 + 2 + 4 + 2 + 4 + 2 // name + null + tableOID + attrNum + typeOID + typeSize + typeMod + format
}
@@ -525,11 +534,11 @@ func (s *PostgreSQLServer) sendRowDescription(session *PostgreSQLSession, column
// Field count
fieldCountBytes := make([]byte, 2)
binary.BigEndian.PutUint16(fieldCountBytes, uint16(len(columns)))
binary.BigEndian.PutUint16(fieldCountBytes, uint16(len(result.Columns)))
msg = append(msg, fieldCountBytes...)
// Field descriptions
for i, col := range columns {
for i, col := range result.Columns {
// Field name
msg = append(msg, []byte(col)...)
msg = append(msg, 0) // null terminator
@@ -544,8 +553,8 @@ func (s *PostgreSQLServer) sendRowDescription(session *PostgreSQLSession, column
binary.BigEndian.PutUint16(attrNum, uint16(i+1))
msg = append(msg, attrNum...)
// Type OID (determine from data)
typeOID := s.getPostgreSQLType(columns, rows, i)
// Type OID (determine from schema if available, fallback to data inference)
typeOID := s.getPostgreSQLTypeFromSchema(result, col, i)
typeOIDBytes := make([]byte, 4)
binary.BigEndian.PutUint32(typeOIDBytes, typeOID)
msg = append(msg, typeOIDBytes...)
@@ -722,8 +731,75 @@ func (s *PostgreSQLServer) getCommandTag(query string, rowCount int) string {
return "SELECT 0"
}
// getPostgreSQLType determines PostgreSQL type OID from data
func (s *PostgreSQLServer) getPostgreSQLType(columns []string, rows [][]sqltypes.Value, colIndex int) uint32 {
// getPostgreSQLTypeFromSchema determines PostgreSQL type OID from schema information first, fallback to data
func (s *PostgreSQLServer) getPostgreSQLTypeFromSchema(result *engine.QueryResult, columnName string, colIndex int) uint32 {
// Try to get type from schema if database and table are available
if result.Database != "" && result.Table != "" {
if tableInfo, err := s.sqlEngine.GetCatalog().GetTableInfo(result.Database, result.Table); err == nil {
if tableInfo.Schema != nil && tableInfo.Schema.RecordType != nil {
// Look for the field in the schema
for _, field := range tableInfo.Schema.RecordType.Fields {
if field.Name == columnName {
return s.mapSchemaTypeToPostgreSQL(field.Type)
}
}
}
}
}
// Handle system columns
switch columnName {
case "_timestamp_ns":
return PG_TYPE_INT8 // PostgreSQL BIGINT for nanosecond timestamps
case "_key":
return PG_TYPE_BYTEA // PostgreSQL BYTEA for binary keys
case "_source":
return PG_TYPE_TEXT // PostgreSQL TEXT for source information
}
// Fallback to data-based inference if schema is not available
return s.getPostgreSQLTypeFromData(result.Columns, result.Rows, colIndex)
}
// mapSchemaTypeToPostgreSQL maps SeaweedFS schema types to PostgreSQL type OIDs
func (s *PostgreSQLServer) mapSchemaTypeToPostgreSQL(fieldType *schema_pb.Type) uint32 {
if fieldType == nil {
return PG_TYPE_TEXT
}
switch kind := fieldType.Kind.(type) {
case *schema_pb.Type_ScalarType:
switch kind.ScalarType {
case schema_pb.ScalarType_BOOL:
return PG_TYPE_BOOL
case schema_pb.ScalarType_INT32:
return PG_TYPE_INT4
case schema_pb.ScalarType_INT64:
return PG_TYPE_INT8
case schema_pb.ScalarType_FLOAT:
return PG_TYPE_FLOAT4
case schema_pb.ScalarType_DOUBLE:
return PG_TYPE_FLOAT8
case schema_pb.ScalarType_BYTES:
return PG_TYPE_BYTEA
case schema_pb.ScalarType_STRING:
return PG_TYPE_TEXT
default:
return PG_TYPE_TEXT
}
case *schema_pb.Type_ListType:
// For list types, we'll represent them as JSON text
return PG_TYPE_JSONB
case *schema_pb.Type_RecordType:
// For nested record types, we'll represent them as JSON text
return PG_TYPE_JSONB
default:
return PG_TYPE_TEXT
}
}
// getPostgreSQLTypeFromData determines PostgreSQL type OID from data (legacy fallback method)
func (s *PostgreSQLServer) getPostgreSQLTypeFromData(columns []string, rows [][]sqltypes.Value, colIndex int) uint32 {
if len(rows) == 0 || colIndex >= len(rows[0]) {
return PG_TYPE_TEXT // Default to text
}

3
weed/server/postgres/server.go
View File

@@ -65,13 +65,16 @@ const (
// PostgreSQL data types
PG_TYPE_BOOL = 16
PG_TYPE_BYTEA = 17
PG_TYPE_INT8 = 20
PG_TYPE_INT4 = 23
PG_TYPE_TEXT = 25
PG_TYPE_FLOAT4 = 700
PG_TYPE_FLOAT8 = 701
PG_TYPE_VARCHAR = 1043
PG_TYPE_TIMESTAMP = 1114
PG_TYPE_JSON = 114
PG_TYPE_JSONB = 3802
// Default values
DEFAULT_POSTGRES_PORT = 5432