[ju-st]

S3 Select is not available anymore for new customers. Athena with columnar file format (eg parquet) in S3 and partitioning with Glue Data catalog is the solution for OP's problem. The cost of this kind of queries is very low because you only pay the actual data consumed/requested. And with the columnar file format Athena only accesses the necessary columns. And the data in the columns is usually compressed so the amount of data is even less.