I would suggest this, 1. Mark the *hiveStore* field in HiveOperator class as depricated. 2. Mark the getters and setters for hiveStore field as depricated, also in javadoc mention that you should use getStore/setStore instead. 3. Update the hiveStore getter setter code to actually set the store of parent class, so we actually don't refer to the hiveStore.
-Priyanka On Mon, May 23, 2016 at 4:09 AM, Yogi Devendra <[email protected] > wrote: > Sending it removing the links for better readability. > > Hi, > > Current implementation of HiveOperator is extended from > AbstractStoreOutputOperator. We have an instance of store in > AbstractStoreOutputOperator. Additionally, HiveOperator declares another > field for store. > > I am thinking of removing this redundant field from HiveOperator. But, this > field is protected; thus, removing it would result in backward > compatibility issue. > > Any recommendations/suggestions on how should I go about this? > > ~ Yogi > > On 23 May 2016 at 15:24, Yogi Devendra <[email protected]> > wrote: > > > Hi, > > > > Current implementation of HiveOperator > > < > https://github.com/apache/incubator-apex-malhar/blob/master/contrib/src/main/java/com/datatorrent/contrib/hive/HiveOperator.java> > is > > extended from AbstractStoreOutputOperator > > < > https://github.com/apache/incubator-apex-malhar/blob/master/library/src/main/java/com/datatorrent/lib/db/AbstractStoreOutputOperator.java > >. > > We have instance of store in AbstractStoreOutputOperator > > < > https://github.com/apache/incubator-apex-malhar/blob/master/library/src/main/java/com/datatorrent/lib/db/AbstractStoreOutputOperator.java#L44 > >. > > Additionally, HiveOperator > > < > https://github.com/apache/incubator-apex-malhar/blob/master/contrib/src/main/java/com/datatorrent/contrib/hive/HiveOperator.java> > declares > > another field for store. > > > > I am thinking of removing this redundant field from HiveOperator. But, > > this field is protected; thus, removing it would result in backward > > compatibility issue. > > Any recommendations/suggestions on how should I go about this? > > > > ~ Yogi > > > > On 12 May 2016 at 11:01, Yogi Devendra <[email protected]> > > wrote: > > > >> Created https://issues.apache.org/jira/browse/APEXMALHAR-2087 for > >> tracking this. > >> > >> ~ Yogi > >> > >> On 11 May 2016 at 22:23, Priyanka Gugale <[email protected]> > >> wrote: > >> > >>> +1, it would be easy to use hive operator if we combine the required > >>> operator together. > >>> > >>> -Priyanka > >>> > >>> On Wed, May 11, 2016 at 3:27 AM, Mohit Jotwani <[email protected]> > >>> wrote: > >>> > >>> > +1 > >>> > > >>> > Regards, > >>> > Mohit > >>> > > >>> > > >>> > > >>> > On Wed, May 11, 2016 at 3:55 PM, Yogi Devendra < > >>> [email protected]> > >>> > wrote: > >>> > > >>> > > Hi, > >>> > > > >>> > > Currently, applications which needs to write to Hive have to > include > >>> 2 > >>> > > operators FSPojoToHiveOperator, HiveOperator from malhar-contrib in > >>> the > >>> > > application DAG. [First operator writes to HDFS; second operator > >>> loads > >>> > > files from HDFS to hive]. > >>> > > > >>> > > I am proposing to wrap these two operators into a > >>> > Module/CompositeOperator. > >>> > > So that, end user can directly include this module into > application. > >>> > > > >>> > > Any feedback? > >>> > > > >>> > > ~ Yogi > >>> > > > >>> > > >>> > >> > >> > > >
