1. Looks like this got added back in by merge mistake. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveBaseResultSet.java Action: revereted. 2. I don't understand this if statement...we do the same thing anyway below? trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveConnection.java Action: reverted. 3. Shouldn't we prevent setAutoCommit(false) rather than setAutoCommit(true)? Hive does not support user-level transactions. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveConnection.java Action: reverted. 4. Where did the number 100 come from? Answer: We needed a non negative number to be returned, not a SQLException. 100 was an abitrary number chosem by the developer who made the change. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveDatabaseMetaData.java 5. This should be a backtick, not an apostrophe. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveDatabaseMetaData.java Action: reverted. 6. This is inaccurate. The database is read/write since you can issue INSERT statements etc. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HiveDatabaseMetaData.java Action: Logic reversed. 7. This should also close the resultSet (if there is one) before nullifying it. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HivePreparedStatement.java Action: result set is closed if it exists. 8. For PreparedStatement, public executeQuery(sql) and others inherited from vanilla Statement should throw, since they aren't supposed to take the SQL string as parameter (that's what prepare is for). Instead, use a protected helper method with a different name (and share with HiveStatement so that excn handling code is reused). trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HivePreparedStatement.java Action: PreparedStatement.executeQuery(String) now throws exception. Protected method PreparedStatement.executeImmediate(sql) was created with implementation, which now includes a call to clearWarnings(). 9. Where is the maxRows enforcement implemented? I may have missed it. Ideally we'd like to turn this into a LIMIT clause for efficiency, but we can do a followup for that. trunk/jdbc/src/java/org/apache/hadoop/hive/jdbc/HivePreparedStatement.java Action: None taken. I did add code to add a limit to "select " statements if max rows > 0. It broke a Junit test where "limit #" was supplied with the SQL statement. So, if max rows is set and limit is supplied in the SQL, what happens? I then reverted the code. Where is the maxRows enforcement implemented? HiveQueryResultSet.next() references it.