Lazy implementation of ResultSet

[alex268]

Current implementation reads all rows to the memory before returning ResultSetImpl. Maybe we need lazy implementation of ResultSet which return rows on demand

[SloNN]

It's a must have feature while working with IDE's like DBeaver. IDE makes a query to database. IDE does not know Database dialect, so it does not apply LIMIT clause on the query. DBeaver executes original query instead and stops reading after reaching expected rows count.

Currently DBeaver makes the query and waiting for results to limit them. Current driver implementation does not return rows to DBeaver until reads all data. So if table is big enough, JDBC Driver reads all data to memory until it receives OOM.

Streaming reading seems to resolve this issue.

[alex268]

Basically, after PR #71 merging, customers can use stream query mode via enabling option 'useStreamResultSets'. So I close this issue. But there is a few improvements and optimizations, like issues #97 and #98