- fallback to limited mode instance tunnel if project is protection
- fix bug when sql statement contains semi-colon
- fix bug of no column info when accessing table from other project
- update odps sdk and fastjson for security reason
- update odps sdk and fastjson for security reason
- add test tool
jdbc_test.sh
- support customizing tunnel endpoint
- add test tool
jdbc_test.sh
with log conf sample
- Support Hive data type.
- Minor update to support hive proxy.
- Support passing
logconffile
to customize logging - Support passing
odps_config
with a file to specify JDBC configuration in order to avoid encoding/decoding the params which contains special characters. Ifodps_config
is passed , neitherURL
norPROP
config will be ignored. - Unify the configuration effects of both
URL_KEY
andPROP_KEY
.PROP_KEY
's priority would be greater thanURL_KEY
's when they both exist. executeUpdate
will return non-negative value.- Add
synchronized
to some methods to protect the safety of concurrent calling.
- Use instance tunnel instead of table tunnel, so that temporary table is no longer created. Demands MaxCompute service >= Sprint27.
- Support set sql properties in
statement.execute()
.
- AutoCommit default is true, and always is true.
- Log for SQLFeatureNotSupportedException on the critical path.
- Refactor the unitest
- Put logview in the warning chain to let be captured in SQLWorkbench/j.
- Remove the hardcode when retrieve version
- Clear javadoc warnings.
- Put all the utils to
/utils
folder.
- Set temp table lifecycle in create SQL (eliminate the round trip cost).
- Accept
VALUES(?, ?, ?)
in prepareStatement. - Reuse one tunnel upload session in
PrepareStatement.executeBatch()
. Commit the session on the close of the PrepareStatement. - Catch the NPE of
instance.getTaskStatus().get("SQL")
. - Support
supportsSchemasInTableDefinitions
to adpat to Kettle.
- Support
isValid()
to adpat to HikariCP. - Support
setReadOnly()
to adapt to HikariCP (only allow for setting it to false). - Add an example to use HikariCP.
- Support
supportsResultSetType()
to adapt to Pentaho. - Support
PreparedStatement.setBytes()
. - Support
getMaxTableNameLength()
to adapt to SQL Workbench/J. - Support
getIdentifierQuoteString()
to adapt to SQL Workbench/J. - Fix the
getUpdateCount()
bug. See Issue #15.
- Support batch insert in PreparedStatement.
- Add an example to migrate data using batch functionality. The example splits the source and uses more than 1 thread to upload.
DatabaseMetaData.getColumns()
Support accoss schema query.- Per-connection logger bug fixed.
- Display uuid in loggers' console handler.
- Use travis to deploy.
user_agent
looks like odps-jdbc-1.1.- Rename OdpsQueryResultSet to ScollResultSet.
- Use project name in log string.
- Throw NPE if
instance.getStatus() == null
.
- SQLWorkbench's wbcopy command supported.
- Compress data when downloading the result set.
- No longer use log4j.
Driver.getParentLogger()
supported.- Per-connection logger (support two connections with different log level.)
- More debug logs for profiling.
- Change the driver name to 'ODPS'.
- Optimize the performance of getting tables.
- No longer supports the concept of catalog (all related functions return null).