Oracle can have both inclusive and exclusive ranges. However, I still have not understood how the test cast will work, if we follow Renato's suggestion to add another parameter to the API method.
Regards, Apos On Fri, Aug 23, 2013 at 2:16 AM, Henry Saputra <henry.sapu...@gmail.com>wrote: > +1 > > Looks like HBase is the one exception? Oracle and Accumulo seem to be > inclusive, and I believe Cassandra also inclusive. > > In the DataStoreTestUtil, we could probably check the type of data store > before executing the delete to pass the right flag. > > - Henry > > > On Thu, Aug 22, 2013 at 3:54 PM, Renato Marroquín Mogrovejo < > renatoj.marroq...@gmail.com> wrote: > > > @Henry, you are right mate (: > > @Apos, you are right as well mate (: but IMHO we could set this value > with > > a default value (the one that most data stores use) and run the tests > like > > this i.e. modifying the tests to make this run well, maybe we could even > > set this new parameter when starting the test so it runs smoothly. > > > > > > Renato M. > > > > > > 2013/8/22 Apostolis Giannakidis <ap.giannaki...@gmail.com> > > > > > Hi Henry, > > > > > > As far as I have understood Renato's proposal, that's correct. > > > But, now that I think of it, if we follow Renato's suggestion, then how > > > will the test case[1] know if it should include the key or not in its > > > checks? > > > > > > [1] > > > > > > > > > https://github.com/apache/gora/blob/trunk/gora-core/src/test/java/org/apache/gora/store/DataStoreTestUtil.java#L747 > > > > > > > > > On Thu, Aug 22, 2013 at 8:08 PM, Henry Saputra < > henry.sapu...@gmail.com > > > >wrote: > > > > > > > HI Renato, > > > > > > > > If the API change include a new parameter to indicate inclusive vs > > > > exclusive then Gora do not have to decide anything and just delegate > > the > > > > new parameter to the corresponding datastore? > > > > > > > > - Henry > > > > > > > > > > > > On Thu, Aug 22, 2013 at 10:43 AM, Renato Marroquín Mogrovejo < > > > > renatoj.marroq...@gmail.com> wrote: > > > > > > > > > Hi all, > > > > > > > > > > I think we could just add an extra parameter to the query API, so > > users > > > > can > > > > > decide programmatically whether they want to use the deletes as > > > inclusive > > > > > or exclusive, and they could do this while programming with Gora's > > API. > > > > And > > > > > we could decide to use a default value for the option that most > data > > > > stores > > > > > support. What do you think? > > > > > > > > > > > > > > > Renato M. > > > > > > > > > > > > > > > 2013/8/18 Apostolis Giannakidis <ap.giannaki...@gmail.com> > > > > > > > > > > > Yes, I can also do both inclusive and exclusive ranges in Oracle > > > NoSQL. > > > > > So > > > > > > it remains to be decided by the Gora API. > > > > > > > > > > > > > > > > > > On Sun, Aug 18, 2013 at 4:06 AM, Scott Stults < > > > > > > sstu...@opensourceconnections.com> wrote: > > > > > > > > > > > > > Thanks for the reply, Apos. Seeing as how this test is in flux > I > > > > won't > > > > > > > worry too much about it now. FWIW, I could do inclusive or > > > exclusive > > > > > > ranges > > > > > > > with Lucene. > > > > > > > > > > > > > > -Scott > > > > > > > > > > > > > > On Aug 17, 2013, at 9:52 PM, Apostolis Giannakidis < > > > > > > > ap.giannaki...@gmail.com> wrote: > > > > > > > > > > > > > > > Hello Scott, > > > > > > > > > > > > > > > > The issue that you just spotted is the same issue that I also > > > > > > > > coincidentally spotted a week ago. > > > > > > > > Keith Turner first identified the issue and documented it in > > > Jira. > > > > > > Please > > > > > > > > see GORA-66. > > > > > > > > https://issues.apache.org/jira/browse/GORA-66 > > > > > > > > > > > > > > > > This is also a blocking issue for me, as it does not allow me > > to > > > > > > complete > > > > > > > > the implementation of deleteByQuery(). Personally, I @Ignored > > > this > > > > > test > > > > > > > > case until GORA-66 is resolved. I saw that the same was done > in > > > > > > Accumulo > > > > > > > > datastore. > > > > > > > > > > > > > > > > I hope this helps, > > > > > > > > Apos > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Sat, Aug 17, 2013 at 8:11 PM, Scott Stults < > > > > > > > > sstu...@opensourceconnections.com> wrote: > > > > > > > > > > > > > > > >> All, > > > > > > > >> > > > > > > > >> I'm having a little trouble getting my head around > > > > deleteByQuery(). > > > > > > The > > > > > > > >> javadoc in the interface indicates that any object that > > matches > > > > the > > > > > > > query > > > > > > > >> should get deleted. The unit test > > > > > > > >> DataStoreTestUtil.testDeleteByQueryFields() expects the > object > > > to > > > > > > still > > > > > > > >> exist with the queried-for fields cleared. To me it seems > like > > > the > > > > > > test > > > > > > > is > > > > > > > >> for an update, rather than a delete. > > > > > > > >> > > > > > > > >> Are my semantics all mixed up? > > > > > > > >> > > > > > > > >> > > > > > > > >> Thanks, > > > > > > > >> -Scott > > > > > > > >> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
