diff --git a/.example/database/gdb/driver/driver/driver.go b/.example/database/gdb/driver/driver/driver.go
index acb2bca93..f672730d4 100644
--- a/.example/database/gdb/driver/driver/driver.go
+++ b/.example/database/gdb/driver/driver/driver.go
@@ -15,9 +15,9 @@ import (
 
 // MyDriver is a custom database driver, which is used for testing only.
 // For simplifying the unit testing case purpose, MyDriver struct inherits the mysql driver
-// gdb.DriverMysql and overwrites its function HandleSqlBeforeCommit.
-// So if there's any sql execution, it goes through MyDriver.HandleSqlBeforeCommit firstly and
-// then gdb.DriverMysql.HandleSqlBeforeCommit.
+// gdb.DriverMysql and overwrites its functions DoQuery and DoExec.
+// So if there's any sql execution, it goes through MyDriver.DoQuery/MyDriver.DoExec firstly
+// and then gdb.DriverMysql.DoQuery/gdb.DriverMysql.DoExec.
 // You can call it sql "HOOK" or "HiJack" as your will.
 type MyDriver struct {
 	*gdb.DriverMysql
diff --git a/database/gdb/gdb_driver_mssql.go b/database/gdb/gdb_driver_mssql.go
index d90ac1f4d..391705a02 100644
--- a/database/gdb/gdb_driver_mssql.go
+++ b/database/gdb/gdb_driver_mssql.go
@@ -71,6 +71,8 @@ func (d *DriverMssql) HandleSqlBeforeCommit(link Link, sql string, args []interf
 	return d.parseSql(str), args
 }
 
+// parseSql does some replacement of the sql before commits it to underlying driver,
+// for support of microsoft sql server.
 func (d *DriverMssql) parseSql(sql string) string {
 	// SELECT * FROM USER WHERE ID=1 LIMIT 1
 	if m, _ := gregex.MatchString(`^SELECT(.+)LIMIT 1$`, sql); len(m) > 1 {
@@ -91,22 +93,20 @@ func (d *DriverMssql) parseSql(sql string) string {
 	index++
 	switch keyword {
 	case "SELECT":
-		// 不含LIMIT关键字则不处理
+		// LIMIT statement checks.
 		if len(res) < 2 ||
 			(strings.HasPrefix(res[index][0], "LIMIT") == false &&
 				strings.HasPrefix(res[index][0], "limit") == false) {
 			break
 		}
-		// 不含LIMIT则不处理
 		if gregex.IsMatchString("((?i)SELECT)(.+)((?i)LIMIT)", sql) == false {
 			break
 		}
-		// 判断SQL中是否含有order by
+		// ORDER BY statement checks.
 		selectStr := ""
 		orderStr := ""
 		haveOrder := gregex.IsMatchString("((?i)SELECT)(.+)((?i)ORDER BY)", sql)
 		if haveOrder {
-			// 取order by 前面的字符串
 			queryExpr, _ := gregex.MatchString("((?i)SELECT)(.+)((?i)ORDER BY)", sql)
 			if len(queryExpr) != 4 ||
 				strings.EqualFold(queryExpr[1], "SELECT") == false ||
@@ -114,8 +114,6 @@ func (d *DriverMssql) parseSql(sql string) string {
 				break
 			}
 			selectStr = queryExpr[2]
-
-			// 取order by表达式的值
 			orderExpr, _ := gregex.MatchString("((?i)ORDER BY)(.+)((?i)LIMIT)", sql)
 			if len(orderExpr) != 4 ||
 				strings.EqualFold(orderExpr[1], "ORDER BY") == false ||
@@ -132,8 +130,6 @@ func (d *DriverMssql) parseSql(sql string) string {
 			}
 			selectStr = queryExpr[2]
 		}
-
-		// 取limit后面的取值范围
 		first, limit := 0, 0
 		for i := 1; i < len(res[index]); i++ {
 			if len(strings.TrimSpace(res[index][i])) == 0 {
@@ -147,7 +143,6 @@ func (d *DriverMssql) parseSql(sql string) string {
 				break
 			}
 		}
-
 		if haveOrder {
 			sql = fmt.Sprintf(
 				"SELECT * FROM "+
diff --git a/database/gdb/gdb_driver_oracle.go b/database/gdb/gdb_driver_oracle.go
index 72eaa5603..d6ccd40fe 100644
--- a/database/gdb/gdb_driver_oracle.go
+++ b/database/gdb/gdb_driver_oracle.go
@@ -75,6 +75,8 @@ func (d *DriverOracle) HandleSqlBeforeCommit(link Link, sql string, args []inter
 	return d.parseSql(str), args
 }
 
+// parseSql does some replacement of the sql before commits it to underlying driver,
+// for support of oracle server.
 func (d *DriverOracle) parseSql(sql string) string {
 	patten := `^\s*(?i)(SELECT)|(LIMIT\s*(\d+)\s*,\s*(\d+))`
 	if gregex.IsMatchString(patten, sql) == false {
@@ -93,22 +95,18 @@ func (d *DriverOracle) parseSql(sql string) string {
 	index++
 	switch keyword {
 	case "SELECT":
-		// 不含LIMIT关键字则不处理
-		if len(res) < 2 || (strings.HasPrefix(res[index][0], "LIMIT") == false && strings.HasPrefix(res[index][0], "limit") == false) {
+		if len(res) < 2 || (strings.HasPrefix(res[index][0], "LIMIT") == false &&
+			strings.HasPrefix(res[index][0], "limit") == false) {
 			break
 		}
-
-		// 取limit前面的字符串
 		if gregex.IsMatchString("((?i)SELECT)(.+)((?i)LIMIT)", sql) == false {
 			break
 		}
-
 		queryExpr, _ := gregex.MatchString("((?i)SELECT)(.+)((?i)LIMIT)", sql)
-		if len(queryExpr) != 4 || strings.EqualFold(queryExpr[1], "SELECT") == false || strings.EqualFold(queryExpr[3], "LIMIT") == false {
+		if len(queryExpr) != 4 || strings.EqualFold(queryExpr[1], "SELECT") == false ||
+			strings.EqualFold(queryExpr[3], "LIMIT") == false {
 			break
 		}
-
-		// 取limit后面的取值范围
 		first, limit := 0, 0
 		for i := 1; i < len(res[index]); i++ {
 			if len(strings.TrimSpace(res[index][i])) == 0 {
@@ -121,10 +119,10 @@ func (d *DriverOracle) parseSql(sql string) string {
 				break
 			}
 		}
-
-		// 也可以使用between,据说这种写法的性能会比between好点,里层SQL中的ROWNUM_ >= limit可以缩小查询后的数据集规模
 		sql = fmt.Sprintf(
-			"SELECT * FROM (SELECT GFORM.*, ROWNUM ROWNUM_ FROM (%s %s) GFORM WHERE ROWNUM <= %d) WHERE ROWNUM_ >= %d",
+			"SELECT * FROM "+
+				"(SELECT GFORM.*, ROWNUM ROWNUM_ FROM (%s %s) GFORM WHERE ROWNUM <= %d)"+
+				" WHERE ROWNUM_ >= %d",
 			queryExpr[1], queryExpr[2], limit, first,
 		)
 	}
diff --git a/database/gdb/gdb_driver_sqlite.go b/database/gdb/gdb_driver_sqlite.go
index 5a17fddeb..c433fcba3 100644
--- a/database/gdb/gdb_driver_sqlite.go
+++ b/database/gdb/gdb_driver_sqlite.go
@@ -59,8 +59,8 @@ func (d *DriverSqlite) GetChars() (charLeft string, charRight string) {
 }
 
 // HandleSqlBeforeCommit deals with the sql string before commits it to underlying sql driver.
-// @todo 需要增加对Save方法的支持，可使用正则来实现替换，
-// @todo 将ON DUPLICATE KEY UPDATE触发器修改为两条SQL语句(INSERT OR IGNORE & UPDATE)
+// TODO 需要增加对Save方法的支持，可使用正则来实现替换，
+// TODO 将ON DUPLICATE KEY UPDATE触发器修改为两条SQL语句(INSERT OR IGNORE & UPDATE)
 func (d *DriverSqlite) HandleSqlBeforeCommit(link Link, sql string, args []interface{}) (string, []interface{}) {
 	return sql, args
 }