JDBC setBoolean implementation

[pedorich-n]

Hi there!
Why does the implementation of setBoolean in JDBC drive converts values to t or f? I get an exception java.sql.SQLException: Query failed : Exception: Cannot compare between BOOLEAN and VARCHAR.

I'm using mapdjdbc-1.0-SNAPSHOT-jar-with-dependencies.jar and as I see, implementation is the same in master branch, https://github.com/omnisci/mapd-core/blob/master/java/mapdjdbc/src/main/java/com/mapd/jdbc/MapDPreparedStatement.java#L167

Example code:

  val sql  = s"select 1 from example where hasException = ?"
  val stmt = ConnectionPool.get().borrow().prepareStatement(sql)
  stmt.setBoolean(1, false)
  val resultSet = stmt.executeQuery()
  println(resultSet.getInt(1))

22:22:24.092 [run-main-6] DEBUG com.mapd.jdbc.MapDStatement - sql is :'select 1 from example where hasException = 'f' LIMIT 100000'
22:22:24.092 [run-main-6] DEBUG com.mapd.jdbc.MapDStatement - afterFnSQL is :'select 1 from example where hasException = 'f' LIMIT 100000'

I obtain connection via ScalikeJDBC's ConnectionPool, but it's the same as DriverManager.getConnection(...)

  Class.forName("com.mapd.jdbc.MapDDriver")
  ConnectionPool.singleton("jdbc:mapd:192.168.1.9:9091:db_test", "mapd", "HyperInteractive")

Am I doing something wrong?

[pedorich-n]

Another question: can I build JDBC driver only? Without building whole core?

[andrewseidl]

Thanks for the report. I've pushed a potential fix here: #276 , should be merged in a day or so.

I've opened #277 requesting adding how to build the JDBC driver on its own. It's possible, but it's currently intertwined with the rest of our CMake-based build. It's basically two thrift -gen java runs to generate the bindings, then need to remember to set a few properties when running maven.