How can I cache dataset at intermediate step of query in spark sql when I have query in string form and not in dataframe API form?

Question

I am getting some inputs from upstream which are sql queries in string form, input file paths and table view name. For execution time improvements on cluste