Release Notes - Tajo - Version 0.9.0

This project has retired. For details please refer to its Attic page.

Changes since Tajo 0.8.0

Sub-task

[TAJO-215] - Catalog should allow compatible types when finding functions
[TAJO-218] - HiveQLAnalyzer has to support cast expression.
[TAJO-517] - Publish Tajo jar to a public maven repository
[TAJO-529] - Fix warnings in tajo-algebra
[TAJO-602] - WorkerResourceManager should be broke down into 3 parts
[TAJO-615] - Implement ADD TABLE RENAME TABLE
[TAJO-659] - Add Tajo JDBC documentation
[TAJO-667] - Add math function documentation
[TAJO-668] - Add datetime function documentation
[TAJO-669] - Add cluster setup documentation
[TAJO-696] - Implement ALTER TABLE ADD COLUMN
[TAJO-697] - Implement ALTER TABLE RENAME COLUMN
[TAJO-736] - Add table management documentation
[TAJO-761] - Implements INTERVAL type
[TAJO-762] - Implements current date/time function
[TAJO-783] - Remove yarn-related code from tajo-core
[TAJO-790] - Implements ADD_MONTHS() function
[TAJO-791] - Implements ADD_DAYS() function
[TAJO-836] - create index support
[TAJO-837] - Register index meta information at Catalog
[TAJO-907] - Implement off-heap tuple block and zero-copy tuple
[TAJO-924] - Merge the current window_function branch to master
[TAJO-991] - Running PullServer on a dedicated JVM process which separates from worker.
[TAJO-992] - Reduce number of hash shuffle output file.
[TAJO-1008] - Protocol buffer De/Serialization for EvalNode
[TAJO-1015] - Add executionblock event in worker
[TAJO-1016] - Refactor worker rpc information
[TAJO-1038] - Remove use of Builder variable in Schema
[TAJO-1060] - Apply updated hadoop versions to README and BUILDING files
[TAJO-1061] - Update build documentation
[TAJO-1062] - Update TSQL documentation
[TAJO-1068] - Add SQL Query documentation
[TAJO-1069] - Add document to explain High Availability support
[TAJO-1077] - Add Derby configuration documentation
[TAJO-1096] - Update download source documentation

Bug

[TAJO-427] - Empty table makes IndexOutOfBoundsException at LEFT OUTER JOIN clause.
[TAJO-563] - INSERT OVERWRITE should not remove data before query success
[TAJO-587] - Query is hanging when OutOfMemoryError occurs in the query master
[TAJO-590] - Rename HiveConverter to HiveQLAnalyzer
[TAJO-619] - SELECT count(1) after joins on text keys causes wrong plans
[TAJO-620] - A join query can cause IndexOutOfBoundsException if one of tables is empty.
[TAJO-624] - Incorrect progress indication
[TAJO-628] - The second stage of distinct aggregation can be scheduled to only one node.
[TAJO-630] - QueryMasterTask never finished when Internal error occurs.
[TAJO-635] - Improve tests of query semantic verification
[TAJO-638] - QueryUnitAttempt causes Invalid event error: TA_UPDATE at TA_ASSIGNED
[TAJO-640] - In inner join clause, empty table can cause a error by order-by clause.
[TAJO-641] - NPE in HCatalogStore.addTable()
[TAJO-645] - Task.Reporter can cause NPE during reporting.
[TAJO-647] - Work unbalance on disk scheduling of DefaultScheduler
[TAJO-648] - TajoWorker does not send correct QM rpc and client rpc ports via heartbeat.
[TAJO-650] - Repartitioner::scheduleHashShuffledFetches should adjust the number of tasks
[TAJO-651] - HcatalogStore should support (de)serialization of RCFile
[TAJO-652] - logical planner cannot handle alias on partition columns
[TAJO-653] - RCFileAppender throws IOException
[TAJO-663] - CREATE TABLE USING RAW doesn't throw ERROR
[TAJO-672] - Wrong progress status when overwrites to partition table
[TAJO-674] - ExplainLogicalPlan can cause NPE when a query includes derived tables
[TAJO-682] - RangePartitionAlgorithm should be improved to handle empty texts
[TAJO-689] - NoSuchElementException occurs during assigning the leaf tasks
[TAJO-690] - infinite loop occurs when rack task is assigning
[TAJO-693] - StatusUpdateTransition in QueryUnitAttempt handles TA_UPDATE incorrectly
[TAJO-698] - Error occurs when FUNCTION and IN statement are used together.
[TAJO-701] - Invalid bytes when creating BlobDatum with offset
[TAJO-705] - CTAS always stores tables with CSV storage type into catalog
[TAJO-706] - In the case of very quick query, client can't get query status.
[TAJO-712] - Fix some bugs after database is supported
[TAJO-713] - Missing INET4 in UniformRangePartition
[TAJO-716] - Using column names actually aliased in aggregation functions can cause planning error.
[TAJO-718] - A group-by clause with the same columns but aliased causes planning error.
[TAJO-719] - JUnit test failures
[TAJO-729] - PreLogicalPlanVerifier verifies distinct aggregation functions incorrectly.
[TAJO-738] - NPE occur when failed in QueryMaster's GlobalPlanner.build().
[TAJO-739] - A subquery with the same column alias caused planning error.
[TAJO-741] - GreedyHeuristicJoinOrderAlgorithm removes some join pairs.
[TAJO-747] - BroadCastJoin omits some data.
[TAJO-748] - Shuffle output numbers of join may be inconsistent.
[TAJO-750] - Join orders affects abnormal to the result data.
[TAJO-754] - failure of INSERT INTO may remove the target table.
[TAJO-759] - Fix findbug errors added recently.
[TAJO-763] - Out of range problem in utc_usec_to()
[TAJO-765] - Incorrect Configuration Classpaths
[TAJO-766] - Test failures in TestExecExternalShellCommand
[TAJO-772] - TajoDump cannot dump upper/lower mixed case database names.
[TAJO-777] - Partition column in function parameter occurs NPE
[TAJO-778] - TPC-DS Q34 occurs NPE
[TAJO-779] - TPC-DS Q46 occurs NPE
[TAJO-786] - TajoDataMetaDatabase::getSchemas creates invalid MetaDataTuple
[TAJO-787] - FilterPushDownRule::visitSubQuery does not consider aliased columns.
[TAJO-792] - Insert table with a qualified target table name can cause error.
[TAJO-795] - PlannerUtil::joinJoinKeyForEachTable need to handle theta-join.
[TAJO-799] - Local query without FROM throws IllegalArgumentException in CLI
[TAJO-800] - CLI's meta command should be aware "TABLE_NAME" style.
[TAJO-802] - No partition columns in WEB catalog page.
[TAJO-803] - INSERT INTO without FROM throws ClassCastException.
[TAJO-805] - Multiple constant in selection emits some columns.
[TAJO-806] - CreateTableNode in CTAS uses a wrong schema as output schema and table schema.
[TAJO-808] - Fix pre-commit build failure
[TAJO-812] - Some methods of TajoDatabaseMetaData should result in an empty tuple list instead of SQLFeatureNotSupportedException.
[TAJO-813] - CLI should support comment character with multi-line query.
[TAJO-816] - NULL delimiter doesn't apply with HCatalogStore
[TAJO-819] - KillQuery does not work query occasionally.
[TAJO-821] - IllegalStateException occurs when a NettyClientBase object is created within single thread.
[TAJO-823] - Missing INET4 handling in DatumFactory.cast()
[TAJO-827] - SUM() overflow in the case of INT4
[TAJO-829] - Same constants in groupby clause may cause planning error.
[TAJO-830] - Some filter conditions with a SUBQUERY are removed by optimizer.
[TAJO-832] - NPE occurs when Exception's message is null in Task.
[TAJO-833] - NPE occurs when using the column as a alias name in the multiple DISTINCT.
[TAJO-848] - PreLogicalPlanVerifier::visitInsert need to find smaller expressions than target columns for a partitioned table.
[TAJO-850] - OUTER JOIN does not properly handle a NULL.
[TAJO-851] - Timestamp type test of TestSQLExpression::testCastFromTable fails in jenkins CI test
[TAJO-852] - Integration test using HCatalog as a catalog store is failed
[TAJO-861] - tajo-dump script are not executable
[TAJO-862] - Restore failure of dumped relations
[TAJO-863] - Column order mismatched in the JOIN query with asterisk selection.
[TAJO-864] - JUnit test failure at TestTimestampDatum.testTimestampConstructor()
[TAJO-866] - COUNT DISTINCT with other aggregation function throws ClassCastException.
[TAJO-867] - OUTER JOIN with empty result subquery produces a wrong result.
[TAJO-868] - TestDateTimeFunctions unit test is occasionally failed
[TAJO-869] - Sometimes, the unit test of testTaskRunnerHistory is failed.
[TAJO-870] - FilterPushDown ignores a partitioned column in CASE expression within WHERE clause
[TAJO-872] - NOW() function has a different value on each task.
[TAJO-873] - Query status is still RUNNING after session expired.
[TAJO-874] - Sometimes InvalidOperationException occurs when aggregates TableStat.
[TAJO-879] - Some data is missing in the case of BROADCAST JOIN and multi-column partition.
[TAJO-880] - NULL in CASE clause occurs Exception.
[TAJO-881] - JOIN with union query occurs NPE
[TAJO-882] - CLI hangs when a error occurs in the GlobalPlanner.
[TAJO-884] - complex join conditions should be supported in ON clause
[TAJO-891] - Complex join conditions with UNION or inline should be supported
[TAJO-894] - Left outer join with partitioned large table and small table returns empty result.
[TAJO-896] - Full outer join query with empty intermediate data doesn't terminate.
[TAJO-897] - PartitionedTableRewriter is repeated several times with same table.
[TAJO-898] - Left outer join with union returns empty result.
[TAJO-899] - Nested now() has different value for each task
[TAJO-902] - Unicode delimiter does not work correctly
[TAJO-904] - ORDER BY Null first support
[TAJO-905] - When to_date() parses some date without day, the result will be wrong.
[TAJO-908] - Fetcher does not retry, when pull server connection was closed
[TAJO-909] - {SortBased, Col}PartitionStoreExec should not write partition keys to files.
[TAJO-912] - Tsql prints wrong version.
[TAJO-913] - Add some missed tests for constant value group-by keys
[TAJO-914] - join queries with constant values can cause schema mismatch in logical plan
[TAJO-916] - SubQuery::computeStatFromTasks occasionally fail.
[TAJO-917] - Using alias name which is same to existing column names causes error
[TAJO-925] - Child ExecutionBlock of JOIN node has different number of shuffle keys.
[TAJO-926] - Join condition including column references of a row-preserving table in left outer join causes incorrect result
[TAJO-927] - Broadcast Join with Large, Small, Large, Small tables makes a wrong plan.
[TAJO-929] - Broadcast join with empty outer join table returns empty result.
[TAJO-934] - Multiple DISTINCT returns null grouping key value.
[TAJO-936] - TestStorages::testSplitable is failed occasionally.
[TAJO-939] - Refactoring the column resolver in LogicalPlan
[TAJO-945] - Connecting to Tajo by JDBC driver failed with SQL Exception "Invalid JDBC URI"
[TAJO-947] - ColPartitionStoreExec can cause URISyntaxException due to special characters
[TAJO-948] - 'INSERT INTO' statement to non existence table casuses NPE.
[TAJO-949] - PullServer does not release files, when a channel throws an internal exception
[TAJO-952] - Wrong default partition volume config
[TAJO-957] - ROUND should be support INT parameter.
[TAJO-960] - TajoCli's problem does not show the current status
[TAJO-961] - TajoCli should exit when at least one query faces error while executing a SQL script.
[TAJO-962] - Column reference used in LIMIT clause incurs NPE.
[TAJO-965] - Upgrade Bytes class and move some methods to others
[TAJO-968] - Self-Join query (including partitioned table) doesn't run unexpectedly using auto broadcast join.
[TAJO-969] - Distributed sort on a large data set may result in incorrect results.
[TAJO-972] - Broadcast join with left outer join returns duplicated rows.
[TAJO-974] - Eliminate unexpected case condition in SubQuery
[TAJO-975] - alias name which is the same to existing column name may cause NPE during PPD
[TAJO-977] - INSERT into a partitioned table as SELECT statement uses a wrong schema.
[TAJO-978] - RoundFloat8 should return Float8Datum type.
[TAJO-979] - Dividing float value by zero should throw "Divide by zero Exception"
[TAJO-980] - execution page in Web UI broken
[TAJO-981] - Help command (\?) in tsql takes too long time.
[TAJO-985] - Client API should be non-blocking
[TAJO-994] - 'count(distinct x)' function counts first null value.
[TAJO-995] - HiveMetaStoreClient wrapper should retry the connection
[TAJO-996] - Sometimes, scheduleFetchesByEvenDistributedVolumes loses some FetchImpls
[TAJO-999] - SequenceFile key class need to be compatible.
[TAJO-1000] - TextDatum.asChar() is incorrect, if client charset is different
[TAJO-1004] - UniformRangePartition cannot deal with unicode ranges
[TAJO-1006] - Fix wrong storage unit for kilo bytes and others.
[TAJO-1009] - A binary eval for column references of the same tables should not be recognized as a join condition
[TAJO-1013] - A complex equality condition including columns of the same table is recognized as a join condition
[TAJO-1017] - TajoConf misuses read & write locks in some functions
[TAJO-1020] - TajoContainerProxy::assignExecutionBlock causes NPE by race condition.
[TAJO-1021] - Remove the member variable Builder from all classes inherited from ProtoObject.
[TAJO-1022] - tsql does not work as background process
[TAJO-1024] - RpcConnectionPool::getConnection can cause NPE at initialization
[TAJO-1025] - Network disconnection during query processing can cause infinite exceptions
[TAJO-1029] - TAJO_PULLSERVER_STANDALONE should be false in default tajo-env.sh
[TAJO-1037] - KillQuery hang in subquery init state
[TAJO-1047] - DefaultTaskScheduler:allocateRackTask is failed occasionally on JDK 1.7
[TAJO-1048] - Missed use of session variables in GlobalPlanner
[TAJO-1050] - RPC client does not retry during connecting
[TAJO-1056] - Wrong resource release or wrong task scheduling
[TAJO-1065] - The \admin -cluster argument doesn't run as expected.
[TAJO-1067] - INSERT OVERWRITE INTO should not remove all partitions.
[TAJO-1072] - CLI gets stuck when wrong host/port is provided
[TAJO-1074] - Query calculates wrong progress before subquery init
[TAJO-1081] - Non-forwarded (simple) query shows wrong rows.
[TAJO-1097] - IllegalArgumentException: RawFileScanner
[TAJO-1098] - LogicalPlanVerifier should validate operations within CASE WHEN clauses.
[TAJO-1099] - LogicalPlanner::convertDataType causes NPE in some cases.
[TAJO-1101] - Broadcast join with a zero-length file table returns wrong result data.
[TAJO-1102] - Self-join with a partitioned table returns wrong result data.
[TAJO-1103] - Insert clause of partitioned table loses some FetchImpls
[TAJO-1104] - Using asterisk with GROUP BY causes NPE.
[TAJO-1106] - Missing session check in getFinishedQuery API
[TAJO-1107] - Broadcast join on non-leaf node scans only first data file.
[TAJO-1110] - JAVA_PULLSERVER_HEAP_MAX in bin/tajo should be increased
[TAJO-1111] - TestKillQuery.testKillQueryFromInitState occasionally fails
[TAJO-1113] - SubQuery in KILLED state should handle unexpected events.

Improvement

[TAJO-153] - Proto(Async|Blocking)RpcClient should retry to connect a server when failed.
[TAJO-196] - Add EngineContext to contain resource information about worker
[TAJO-356] - Improve TajoClient to directly get query results in the first request
[TAJO-425] - RAWFILE_SYNC_INTERVAL has not default value.
[TAJO-589] - Add fine grained progress indicator for each task
[TAJO-614] - Explaining a logical node should use ExplainLogicalPlanVisitor.
[TAJO-616] - SequenceFile support
[TAJO-617] - Rename BIN/tajo_dump BIN/tajo-dump
[TAJO-634] - ExecutionBlock must be sorted by start time in querydetail.jsp
[TAJO-644] - Support quoted identifiers
[TAJO-654] - Separate TajoWorker into yarn worker and standby worker
[TAJO-662] - The tasks of CTAS on a partitioned table should be fine grained
[TAJO-665] - sort buffer size must be dealt as long type values.
[TAJO-670] - Change daemon's hostname to canonical hostname
[TAJO-673] - Assign proper number of tasks when inserting into partitioned table
[TAJO-675] - maximum frame size of frameDecoder should be increased
[TAJO-691] - HashJoin or HashAggregation is too slow if there is many unique keys
[TAJO-699] - Create a table using LIKE
[TAJO-709] - Add .reviewboardrc and use rbt instead of post-review
[TAJO-714] - Enable setting Parquet tuning parameters
[TAJO-715] - hadoop version upgrade to 2.3.0
[TAJO-717] - Improve file splitting for large number of splits
[TAJO-725] - Broadcast JOIN should supports multiple tables
[TAJO-728] - Supports expressions in 'IN predicate'
[TAJO-732] - Support executing LINUX shell command and HDFS command.
[TAJO-734] - Arrange TajoCli output message.
[TAJO-735] - Remove multiple SLF4J bindings message.
[TAJO-737] - Change version message when daemon starts up.
[TAJO-743] - Change the default resource allocation policy of leaf tasks
[TAJO-745] - APIs in TajoClient and JDBC should be case sensitive.
[TAJO-755] - ALTER TABLESPACE LOCATION support
[TAJO-758] - Supports parameter values in the SQL file.
[TAJO-768] - Improve the log4j configuration
[TAJO-769] - A minor improvements for HCatalogStore
[TAJO-789] - Improve shuffle URI
[TAJO-793] - CLI should be able to exit when single query is failed.
[TAJO-797] - Implicit type conversion support
[TAJO-801] - Multiple distinct should be supported.
[TAJO-804] - Bump up Parquet version to 1.4.2
[TAJO-807] - Implement Round(numeric, int) function.
[TAJO-811] - add simple fifo scheduler support
[TAJO-824] - Improve SimpleParser to handle JSON statements
[TAJO-840] - Improve query result print with counting empty table.
[TAJO-842] - NULL handling in JDBC.
[TAJO-843] - implements COALESCE for BOOLEAN, DATE, TIME, TIMESTAMP
[TAJO-844] - JDBC should be support getTime, getDate, and getTimestamp.
[TAJO-846] - Clean up the task history in woker
[TAJO-853] - Refactoring FilterPushDown for OUTER JOIN
[TAJO-895] - ConstEval should not be included in target list of projectable nodes
[TAJO-900] - Reducing memory usage during query processing
[TAJO-903] - Some left outer join cases are not optimized as the broadcast join.
[TAJO-906] - Runtime code generation for evaluating expression trees
[TAJO-910] - Simple query (non-forwarded query) should be supported against partition tables.
[TAJO-911] - Refactoring Mysql/Maria Catalog Store
[TAJO-928] - Session variables should override query configs in TajoConf.
[TAJO-931] - Output file can be punctuated depending on the file size.
[TAJO-932] - Upgrade Parquet to 1.5.0.
[TAJO-933] - Fork some classes of Parquet as builtin third-party classes
[TAJO-937] - Should use tajo.util.VersionInfo instead of TajoConstants.TAJO_VERSION
[TAJO-953] - RawFile should release a DirectBuffer immediately
[TAJO-956] - CONCAT should be support multiple params and null param.
[TAJO-966] - Range partition should support split of multiple characters.
[TAJO-983] - Worker should directly read Intermediate data stored in localhost rather than fetching
[TAJO-984] - Improve the default data type handling in RowStoreUtil
[TAJO-987] - Hash shuffle should be balanced according to intermediate volumes
[TAJO-989] - Cleanup of child blocks after parent execution block is complete
[TAJO-990] - Implement a tool to find tajo configurations.
[TAJO-1010] - Improve multiple DISTINCT aggregation.
[TAJO-1027] - Upgrade Hive to 0.13.0 and 0.13.1
[TAJO-1028] - JDBC should support SET command.
[TAJO-1030] - Not supported JDBC APIs should return empty results instead of Exception
[TAJO-1034] - Reduce Explicit Use of JVM Internal Class
[TAJO-1040] - Misuse netty HashedWheelTimer
[TAJO-1046] - Remove hadoop native dependency of pullserver
[TAJO-1049] - Remove the parallel degree limit up to the maximum cluster capacity
[TAJO-1052] - (Umbrella) Add and Update user documentation for 0.9.0
[TAJO-1071] - should be possible to get long query results with no prompt
[TAJO-1093] - DateTimeFormat.to_char() is slower than SimpleDateFormat.format()

New Feature

[TAJO-20] - INSERT INTO ... SELECT
[TAJO-30] - Parquet Integration
[TAJO-353] - Add Database support to Tajo
[TAJO-377] - Implement concat function
[TAJO-378] - Implement concat_ws function.
[TAJO-480] - Umbrella Jira for adding ALTER TABLE statement
[TAJO-704] - TajoMaster HA
[TAJO-711] - Add Avro storage support
[TAJO-847] - Supporting MariaDB-based Store, which is compatible with MySQL.
[TAJO-849] - Add Parquet storage to HCatalogStore
[TAJO-860] - Implements TRUNCATE table.
[TAJO-1105] - Add thread which detects JVM pauses like HADOOP's

Task

[TAJO-605] - Rename Options to KeyValueList
[TAJO-621] - Add DOAP file for Tajo
[TAJO-632] - add intellij idea projects files into git ignore
[TAJO-642] - Change tajo documentation tool to sphinx
[TAJO-657] - Missing table stat in RCFile
[TAJO-681] - Embed sphinx rtd theme into tajo-docs
[TAJO-694] - Bump up hadoop to 2.3.0
[TAJO-700] - Update site, wikis, pom.xml and other resources to point to the new repository location
[TAJO-730] - Update Tajo site to reflect graduation
[TAJO-752] - Escalate sub modules in tajo-core into the top-level modules
[TAJO-753] - Clean up of maven dependencies
[TAJO-788] - Update Tajo documentation and README, and BUILDING
[TAJO-810] - Update Tajo site for 0.8.0 release
[TAJO-814] - Set up Travis CI builds
[TAJO-817] - tajo-core should not skip deploy.
[TAJO-820] - Add missing license header to 0.8.0 release announcement.
[TAJO-834] - Add Travis notification to issues@tajo.a.o and IRC.
[TAJO-859] - Update site for new committer Alvin Henrick
[TAJO-886] - Add IRC page to community section in site.
[TAJO-887] - Eliminate HiveQL support feature
[TAJO-890] - Redirect stdout of maven test to /dev/null in Travis CI script
[TAJO-1001] - Add missed postgresql license to NOTICE.txt and LICENSE.txt
[TAJO-1007] - Update site for new committer and new contributors
[TAJO-1054] - Wrong comment in ByteUtils.splitWorker()
[TAJO-1070] - BSTIndexScanExec should not seek a negative offset
[TAJO-1078] - Update contributor list

Test

[TAJO-686] - Integration test aborted