parser: Allows no target cols to be specified in IMPORT INTO query.

This is an add on to cockroachdb#39023. It adds the grammar required to support
queries such as:
`IMPORT INTO test CSV DATA ('...');`

No target columns implies that all columns are imported into from the
data sources.

Release note: None

docs/generated/sql/bnf/import_csv.bnf

@@ -5,3 +5,5 @@ import_stmt ::=
 	| 'IMPORT' 'TABLE' table_name '(' table_elem_list ')' 'CSV' 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 
 	| 'IMPORT' 'INTO' table_name '(' insert_column_list ')' 'CSV' 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 'WITH' kv_option_list
 	| 'IMPORT' 'INTO' table_name '(' insert_column_list ')' 'CSV' 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 
+	| 'IMPORT' 'INTO' table_name 'CSV' 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 'WITH' kv_option_list
+	| 'IMPORT' 'INTO' table_name 'CSV' 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 

docs/generated/sql/bnf/import_dump.bnf

@@ -5,3 +5,5 @@ import_stmt ::=
 	| 'IMPORT' 'TABLE' table_name 'FROM' import_format file_location 
 	| 'IMPORT' 'INTO' table_name '(' insert_column_list ')' import_format 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 'WITH' kv_option_list
 	| 'IMPORT' 'INTO' table_name '(' insert_column_list ')' import_format 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 
+	| 'IMPORT' 'INTO' table_name import_format 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 'WITH' kv_option_list
+	| 'IMPORT' 'INTO' table_name import_format 'DATA' '(' file_location ( ( ',' file_location ) )* ')' 

docs/generated/sql/bnf/stmt_block.bnf

@@ -130,6 +130,7 @@ import_stmt ::=
 	| 'IMPORT' 'TABLE' table_name 'CREATE' 'USING' string_or_placeholder import_format 'DATA' '(' string_or_placeholder_list ')' opt_with_options
 	| 'IMPORT' 'TABLE' table_name '(' table_elem_list ')' import_format 'DATA' '(' string_or_placeholder_list ')' opt_with_options
 	| 'IMPORT' 'INTO' table_name '(' insert_column_list ')' import_format 'DATA' '(' string_or_placeholder_list ')' opt_with_options
+	| 'IMPORT' 'INTO' table_name import_format 'DATA' '(' string_or_placeholder_list ')' opt_with_options
 
 insert_stmt ::=
 	opt_with_clause 'INSERT' 'INTO' insert_target insert_rest returning_clause

pkg/ccl/importccl/import_stmt_test.go

@@ -1739,8 +1739,44 @@ func TestImportIntoCSV(t *testing.T) {
 		sqlDB.Exec(t, "DROP DATABASE targetcols")
 	})
 
-	// TODO(adityamaru): Add test for IMPORT INTO without target columns specified
-	// once grammar has been added.
+	// Tests IMPORT INTO without any target columns specified. This implies an
+	// import of all columns in the exisiting table.
+	t.Run("no-target-cols-specified", func(t *testing.T) {
+		sqlDB.Exec(t, "CREATE DATABASE targetcols; USE targetcols")
+		sqlDB.Exec(t, `CREATE TABLE t (a INT, b STRING)`)
+
+		// Insert the test data
+		insert := []string{"''", "'text'", "'a'", "'e'", "'l'", "'t'", "'z'"}
+
+		if tx, err := db.Begin(); err != nil {
+			t.Fatal(err)
+		} else {
+			for i, v := range insert {
+				sqlDB.Exec(t, fmt.Sprintf("INSERT INTO t (a, b) VALUES (%d, %s)", i, v))
+			}
+
+			if err := tx.Commit(); err != nil {
+				t.Fatal(err)
+			}
+		}
+
+		sqlDB.Exec(t, fmt.Sprintf("IMPORT INTO t CSV DATA (%s)", testFiles.files[0]))
+
+		var result int
+		numExistingRows := len(insert)
+		// Verify that all columns have been populated with imported data.
+		sqlDB.QueryRow(t, `SELECT count(*) FROM t WHERE a IS NOT NULL`).Scan(&result)
+		if expect := numExistingRows + rowsPerFile; result != expect {
+			t.Fatalf("expected %d rows, got %d", expect, result)
+		}
+
+		sqlDB.QueryRow(t, `SELECT count(*) FROM t WHERE b IS NOT NULL`).Scan(&result)
+		if expect := numExistingRows + rowsPerFile; result != expect {
+			t.Fatalf("expected %d rows, got %d", expect, result)
+		}
+
+		sqlDB.Exec(t, "DROP DATABASE targetcols")
+	})
 }
 
 func BenchmarkImport(b *testing.B) {

pkg/sql/parser/parse_test.go

@@ -1294,6 +1294,7 @@ func TestParse(t *testing.T) {
 		{`IMPORT TABLE foo (id INT8, email STRING, age INT8) CSV DATA ('path/to/some/file', $1) WITH comma = ',', "nullif" = 'n/a', temp = $2`},
 		{`IMPORT TABLE foo FROM PGDUMPCREATE 'nodelocal:///foo/bar' WITH temp = 'path/to/temp'`},
 		{`IMPORT INTO foo(id, email) CSV DATA ('path/to/some/file', $1) WITH temp = 'path/to/temp'`},
+		{`IMPORT INTO foo CSV DATA ('path/to/some/file', $1) WITH temp = 'path/to/temp'`},
 
 		{`IMPORT PGDUMP 'nodelocal:///foo/bar' WITH temp = 'path/to/temp'`},
 		{`EXPLAIN IMPORT PGDUMP 'nodelocal:///foo/bar' WITH temp = 'path/to/temp'`},

pkg/sql/parser/sql.y

@@ -1839,6 +1839,11 @@ import_stmt:
     name := $3.unresolvedObjectName().ToTableName()
     $$.val = &tree.Import{Table: &name, Into: true, IntoCols: $5.nameList(), FileFormat: $7, Files: $10.exprs(), Options: $12.kvOptions()}
   }
+| IMPORT INTO table_name import_format DATA '(' string_or_placeholder_list ')' opt_with_options
+  {
+    name := $3.unresolvedObjectName().ToTableName()
+    $$.val = &tree.Import{Table: &name, Into: true, IntoCols: nil, FileFormat: $4, Files: $7.exprs(), Options: $9.kvOptions()}
+  }
 | IMPORT error // SHOW HELP: IMPORT
 
 // %Help: EXPORT - export data to file in a distributed manner

pkg/sql/sem/tree/import.go

@@ -46,6 +46,8 @@ func (node *Import) Format(ctx *FmtCtx) {
 				ctx.WriteByte('(')
 				ctx.FormatNode(&node.IntoCols)
 				ctx.WriteString(") ")
+			} else {
+				ctx.WriteString(" ")
 			}
 		} else {
 			ctx.WriteString("TABLE ")