Major Description
Spark SQL uses the class Origin for tracking the position of each TreeNode in the SQL query text. When there is a parser error, we can show the position info in the error message:
> sql("create tabe foo(i int)") org.apache.spark.sql.catalyst.parser.ParseException: no viable alternative at input 'create tabe'(line 1, pos 7) == SQL == create tabe foo(i int) -------^^^
It contains two fields: line and startPosition. This is enough for the parser since the SQL query text is known.
However, the SQL query text is unknown in the execution phase. One solution is to include the query text in Origin and show it in the runtime error message.
