| Fixed issue where the c.t.TupleEntryCollector was not set to "null" on the c.o.OperationCall before c.o.Operation#cleanup() was called to prevent the method from emitting values during cleanup. See Operation#flush(). |
|
More...
|
almost 13 years ago
|
| use version release string |
|
More...
|
almost 13 years ago
|
| Fixed issue where c.f.h.ProcessFlow would not honor c.f.FlowListener instances. Currently does not support the #onThrowable event. |
|
More...
|
almost 13 years ago
|
| Updated c.p.a.Unique to use c.o.b.FirstNBuffer to improve performance. Added c.o.b.FirstNBuffer to provide a faster implementation of returning the first N tuples encountered in a grouping. |
|
More...
|
almost 13 years ago
|
| Updated junit to version 4.11. |
|
More...
|
almost 13 years ago
|
| Update default Apache Hadoop support to version 1.1.x. End support for 0.20.2. |
|
More...
|
almost 13 years ago
|
| fix for missing slf4j from test classpath |
|
More...
|
almost 13 years ago
|
| Updated c.f.FlowDef to accept classpath elements that allow for pipe assemblies to load additional resources from the current context j.l.ClassLoader. |
|
More...
|
almost 13 years ago
|
| updated javadoc |
|
More...
|
almost 13 years ago
|
| downgrade info logs to debug |
|
More...
|
almost 13 years ago
|
| Updated error messages in c.t.Fields, delegate property initialization to c.f.Flow sub-classes. @fderose |
|
More...
|
almost 13 years ago
|
| include missing scheme in fully qualified identifier |
|
More...
|
almost 13 years ago
|
| improved planner error message when Every follows HashJoin |
|
More...
|
about 13 years ago
|
| change property platform.includes to test.platform.includes to retain 2.0 compatibility |
|
More...
|
about 13 years ago
|
| improved trace functionality, added tests |
|
More...
|
about 13 years ago
|
| simplified uuid code, removed dead methods, lazy id for pipes |
|
More...
|
about 13 years ago
|
| fix javadoc |
|
More...
|
about 13 years ago
|
| point javadoc to Hadoop 1.0.4 |
|
More...
|
about 13 years ago
|
| fix for npe when child sub-assembly did not set previous |
|
More...
|
about 13 years ago
|
| Removed Hadoop oro dependency from build and test runtime classpaths to stop transient build failures. |
|
More...
|
about 13 years ago
|
| Added ability to pass System level properties into platform level property sets to override defaults during testing. |
|
More...
|
about 13 years ago
|
| Fixed issue where c.t.l.FileTap#getFullIdentifier() was not returning the fully qualified path. |
|
More...
|
about 13 years ago
|
| Added c.t.h.HfsProps to localize optional Hadoop HDFS specific properties, specifically provides properties for enabling the combining of small files into larger splits. Updated c.t.h.Hfs to allow for smaller files to be combined into fewer splits, thus fewer map tasks. @sjlee |
|
More...
|
about 13 years ago
|
| Updated c.p.SubAssembly to support setting local and step properties via the c.p.ConfigDef. |
|
More...
|
about 13 years ago
|
| revised javadoc |
|
More...
|
about 13 years ago
|
| Updated c.o.Buffer to allow implementations to disable nulling of non-grouping fields after the arguments iterator has completed. This simplifies appending aggregated fields to the incoming tuple stream. |
|
More...
|
about 13 years ago
|
| cleaned up test platform methods, improved error message on TupleEntryOutputCollector |
|
More...
|
about 13 years ago
|
| update guava to 14.0.1 |
|
More...
|
about 13 years ago
|
| set argument fields in test concrete call |
|
More...
|
about 13 years ago
|
| merge 2.1.6 into 2.2 |
|
More...
|
about 13 years ago
|
