parser: move LOAD DATA REMOTE into LOAD DATA (#41091)

ref #40499

executor/load_data.go

@@ -28,6 +28,7 @@ import (
 
 	"github.com/pingcap/errors"
 	"github.com/pingcap/failpoint"
+	backup "github.com/pingcap/kvproto/pkg/brpb"
 	"github.com/pingcap/tidb/br/pkg/storage"
 	"github.com/pingcap/tidb/expression"
 	"github.com/pingcap/tidb/parser/ast"
@@ -63,10 +64,6 @@ type LoadDataExec struct {
 // Next implements the Executor Next interface.
 func (e *LoadDataExec) Next(ctx context.Context, req *chunk.Chunk) error {
 	req.GrowAndReset(e.maxChunkSize)
-	// TODO: support load data without local field.
-	if e.FileLocRef == ast.FileLocServer {
-		return errors.New("Load Data: don't support load data without local field")
-	}
 	// TODO: support lines terminated is "".
 	if len(e.loadDataInfo.LinesInfo.Terminated) == 0 {
 		return errors.New("Load Data: don't support load data terminated is nil")
@@ -79,8 +76,21 @@ func (e *LoadDataExec) Next(ctx context.Context, req *chunk.Chunk) error {
 	}
 
 	switch e.FileLocRef {
-	case ast.FileLocServer:
-		panic("FileLocServer should be handled earlier")
+	case ast.FileLocServerOrRemote:
+		u, err := storage.ParseRawURL(e.loadDataInfo.Path)
+		if err != nil {
+			return err
+		}
+		var filename string
+		u.Path, filename = filepath.Split(u.Path)
+		b, err := storage.ParseBackendFromURL(u, nil)
+		if err != nil {
+			return err
+		}
+		if b.GetLocal() != nil {
+			return errors.Errorf("Load Data: don't support load data from tidb-server's disk")
+		}
+		return e.loadFromRemote(ctx, b, filename)
 	case ast.FileLocClient:
 		// let caller use handleQuerySpecial to read data in this connection
 		sctx := e.loadDataInfo.ctx
@@ -90,27 +100,15 @@ func (e *LoadDataExec) Next(ctx context.Context, req *chunk.Chunk) error {
 			return errors.New("Load Data: previous load data option wasn't closed normally")
 		}
 		sctx.SetValue(LoadDataVarKey, e.loadDataInfo)
-	case ast.FileLocRemote:
-		return e.loadFromRemote(ctx)
 	}
 	return nil
 }
 
-func (e *LoadDataExec) loadFromRemote(ctx context.Context) error {
-	u, err := storage.ParseRawURL(e.loadDataInfo.Path)
-	if err != nil {
-		return err
-	}
-	var filename string
-	u.Path, filename = filepath.Split(u.Path)
-	b, err := storage.ParseBackendFromURL(u, nil)
-	if err != nil {
-		return err
-	}
-	if b.GetLocal() != nil {
-		return errors.Errorf("Load Data: don't support load data from tidb-server when set REMOTE, path %s", e.loadDataInfo.Path)
-	}
-
+func (e *LoadDataExec) loadFromRemote(
+	ctx context.Context,
+	b *backup.StorageBackend,
+	filename string,
+) error {
 	opt := &storage.ExternalStorageOptions{}
 	if InTest {
 		opt.NoCredentials = true

executor/loadremotetest/one_csv_test.go

@@ -26,7 +26,7 @@ func (s *mockGCSSuite) TestLoadCSV() {
 	s.tk.MustExec("CREATE TABLE load_csv.t (i INT, s varchar(32));")
 
 	// no-new-line-at-end
-	sql := fmt.Sprintf(`LOAD DATA REMOTE INFILE 'gcs://test-bucket/no-new-line-at-end.csv?endpoint=%s' INTO TABLE load_csv.t
+	sql := fmt.Sprintf(`LOAD DATA INFILE 'gcs://test-bucket/no-new-line-at-end.csv?endpoint=%s' INTO TABLE load_csv.t
 		FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'
 		LINES TERMINATED BY '\n' IGNORE 1 LINES;`, gcsEndpoint)
 	s.tk.MustExec(sql)
@@ -39,7 +39,7 @@ func (s *mockGCSSuite) TestLoadCSV() {
 	s.tk.MustExec("TRUNCATE TABLE load_csv.t;")
 
 	// new-line-at-end
-	sql = fmt.Sprintf(`LOAD DATA REMOTE INFILE 'gcs://test-bucket/new-line-at-end.csv?endpoint=%s' INTO TABLE load_csv.t
+	sql = fmt.Sprintf(`LOAD DATA INFILE 'gcs://test-bucket/new-line-at-end.csv?endpoint=%s' INTO TABLE load_csv.t
 		FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'
 		LINES TERMINATED BY '\n' IGNORE 1 LINES;`, gcsEndpoint)
 	s.tk.MustExec(sql)
@@ -52,6 +52,6 @@ func (s *mockGCSSuite) TestLoadCSV() {
 	s.tk.MustExec("TRUNCATE TABLE load_csv.t;")
 
 	// can't read file at tidb-server
-	sql = "LOAD DATA REMOTE INFILE '/etc/passwd' INTO TABLE load_csv.t;"
+	sql = "LOAD DATA INFILE '/etc/passwd' INTO TABLE load_csv.t;"
 	s.tk.MustContainErrMsg(sql, "don't support load data from tidb-server")
 }

parser/ast/dml.go

@@ -1804,14 +1804,12 @@ func (n *ColumnNameOrUserVar) Accept(v Visitor) (node Node, ok bool) {
 type FileLocRefTp int
 
 const (
-	// FileLocServer is used when there's no keywords in SQL, which means the data file should be located on the tidb-server.
-	FileLocServer FileLocRefTp = iota
+	// FileLocServerOrRemote is used when there's no keywords in SQL, which means the data file should be located on the
+	// tidb-server or on remote storage (S3 for example).
+	FileLocServerOrRemote FileLocRefTp = iota
 	// FileLocClient is used when there's LOCAL keyword in SQL, which means the data file should be located on the MySQL
 	// client.
 	FileLocClient
-	// FileLocRemote is used when there's REMOTE keyword in SQL, which means the data file should be located on a remote
-	// server, such as a cloud storage.
-	FileLocRemote
 )
 
 // LoadDataStmt is a statement to load data from a specified file, then insert this rows into an existing table.
@@ -1837,11 +1835,9 @@ type LoadDataStmt struct {
 func (n *LoadDataStmt) Restore(ctx *format.RestoreCtx) error {
 	ctx.WriteKeyWord("LOAD DATA ")
 	switch n.FileLocRef {
-	case FileLocServer:
+	case FileLocServerOrRemote:
 	case FileLocClient:
 		ctx.WriteKeyWord("LOCAL ")
-	case FileLocRemote:
-		ctx.WriteKeyWord("REMOTE ")
 	}
 	ctx.WriteKeyWord("INFILE ")
 	ctx.WriteString(n.Path)

parser/misc.go

@@ -588,7 +588,6 @@ var tokenMap = map[string]int{
 	"RELEASE":                  release,
 	"RELOAD":                   reload,
 	"REMOVE":                   remove,
-	"REMOTE":                   remote,
 	"RENAME":                   rename,
 	"REORGANIZE":               reorganize,
 	"REPAIR":                   repair,