Anonymous
Not applicable

I would say don't use spark for data that has awful/no schemas. Use spark for scale and data with a schema. Maybe try to fix whatever is creating these messages.