sky 发表于 2015-3-5 03:28:18

Hive 1.0 正式版发布,数据仓库平台

3月21日 深圳 OSC 源创会开始报名罗,送华为海思开发板
  Apache Hive 1.0.0 正式发布了。该版本原本是要命名为 Hive 0.14.1,但是团队感觉到了该用 1.x.y 的方式命名的时候了。不过该版本改变的内容并不多,值得关注的有两个:

[*]  为 HiveMetaStoreClient 定义 API
[*]  移除 HiveServer 1,全面使用 HiveServer 2
  Hive 1.0 版本完整的改进记录包括:
Bug

[*]   - Index creation on a skew table fails
[*]   - Drop cascade database fails when the db has any tables with indexes
[*]   - ORC file and struct column names are case sensitive
[*]   - HS2 & Metastore using SASL out of memory error when curl sends a get request
[*]   - SerDe Properties are not considered by show create table Command
[*]   - IN operator for partition column fails when the partition column type is DATE
[*]   - Add batch retrieve partition objects for metastore direct sql
[*]   - schematool fails on Postgres versions < 9.2
[*]   - HMS on Oracle incompatibility
[*]   - Table statistic collection on counter failed due to table name character case.
[*]   - Hive 14 upgrade scripts can fail for statistics if database was created using auto-create
[*]   - handle Derby and Oracle errors with joins and filters in Direct SQL in a invalid-DB-specific path
[*]   - Querying partition does not work with JDO enabled against PostgreSQL
[*]   - CBO skipped due to SemanticException: Line 0:-1 Both left and right aliases encountered in JOIN 'avg_cs_ext_discount_amt'
[*]   - Obsolete default values in webhcat-default.xml
[*]   - Dynamic partition pruning can result in NPE during query compilation
[*]   - Remove SSLv2Hello from list of disabled protocols
[*]   - hcatalog process don't exit because of non daemon thread
[*]   - Switch to Tez 0.5.2
[*]   - Vectorization on partitioned table throws ArrayIndexOutOfBoundsException when partitions are not of same #of columns
[*]   - errors when selecting a struct field within an array from ORC based tables
[*]   - Switch to calcite 0.9.2
[*]   - incorrect upgrade script for Oracle (13->14)
[*]   - non-synchronized access to split list in OrcInputFormat
[*]   - Some Vectorized String CONCAT expressions result in runtime error Vectorization: Unsuported vector output type: StringGroup
[*]   - Mapjoin with LateralViewJoin generates wrong plan in Tez
[*]   - HIVE-8876 also affects Postgres < 9.2
[*]   - Delta files created by hive hcatalog streaming cannot be compacted
[*]   - Vectorized IF expr broken for the scalar and scalar case
[*]   - join38.q (without map join) produces incorrect result when testing with multiple reducers
[*]   - Join tests fail on Tez
[*]   - TezJobMonitor in-place updates logs too often to logfile
[*]   - select constant in union all followed by group by gives wrong result
[*]   - OrcFileMergeOperator may create merge file that does not match properties of input files
[*]   - Rename "Tez File Merge Work" to smaller name
[*]   - Fix for HIVE-8735 is incorrect (stats with long paths)
[*]   - Potential NPE in OrcStruct for list and map types
[*]   - Query may generate different results depending on the number of reducers
[*]   - union all query in cbo test has undefined ordering
[*]   - Backport HIVE-8827 (Remove SSLv2Hello from list of disabled protocols) to 0.14 branch
[*]   - HiveOnTez: mix of union all, distinct, group by generates error
[*]   - HIVE_LOCKS uses int instead of bigint hive-txn-schema-0.14.0.mssql.sql
[*]   - stats19 test is environment-dependant
[*]   - Place an upper bound for SARG CNF conversion
[*]   - Vectorized Coalesce for strings is broken
[*]   - Change default tez install directory to use /tmp instead of /user and create the directory if it does not exist
[*]   - HiveServer2 leaks FileSystem objects in FileSystem.CACHE
[*]  - java.lang.ClassCastException: org.apache.hadoop.hive.serde2.io.HiveVarcharWritable cannot be cast to org.apache.hadoop.hive.common.type.HiveVarchar when joining tables
[*]   - Cached expression feature broken in one case
[*]   - move Microsoft copyright to NOTICE file
[*]   - Export of a large table causes OOM in Metastore and Client
[*]   - Intermittent NPE in SessionHiveMetaStoreClient.alterTempTable
[*]   - Enhance retry logic wrt DB access in TxnHandler
[*]   - Backport: Fastpath for limited fetches from unpartitioned tables
[*]   - NPE in org.apache.hadoop.hive.metastore.txn.TxnHandler.determineDatabaseProduct()
[*]   - RetryingMetaStoreClient does not retry JDOExceptions
[*]   - sql std auth should disallow built-in udfs that allow any java methods to be called
[*]   - schematool is broken in hive 1.0.0
Improvement

[*]   - Make HiveMetaStoreClient a public API
[*]
   - Check>
Task

[*]   - Delete Hiveserver1
页: [1]
查看完整版本: Hive 1.0 正式版发布,数据仓库平台