SQL: Support of creating tables from multiple paths

Spark support loading data frames from multiple paths - https://spark.apache.org/docs/2.0.2/api/java/org/apache/spark/sql/DataFrameReader.html#load(scala.collection.Seq)

When I tried to emulate similar with Snappy SQL, I got really weird results:

CREATE TEMPORARY TABLE users(uuid string, date_created long)
 USING parquet OPTIONS(
   path 's3a://bucket/users_2018_01_01-2018_01_31/',
   path 's3a://bucket/users_2018_02_01-2018_02_31/'));

The query succeeds, but when I select data it contains all data from the last path (path 's3a://bucket/users_2018_02_01-2018_02_31/') without data from the first path. 

Unfortunately, I cannot use hive partitioning here to add data by partitions.