Re: DocValues only faceting tests in distributed mode.
Thanks Ishan! I worked out a different approach with the minimal mutable schema stuff, we'll see if it makes sense. On Fri, Nov 25, 2016 at 3:23 PM, Ishan Chattopadhyaya wrote: > The test doesn't have anything to do with faceting, though. Wasn't sure that > you were specifically looking for a faceting test with docValues=true, > indexed=false, or just a general one. > > On Sat, Nov 26, 2016 at 4:48 AM, Ishan Chattopadhyaya > wrote: >> >> Hi Erick, >> As part of working on SOLR-5944, I wrote a few tests with docValues=true, >> indexed=false, stored=false. You can find it in the latest patch there. The >> schema file is called: schema-inplace-updates.xml, and one of the tests >> using that schema is: TestInPlaceUpdatesDistrib >> Regards, >> Ishan >> >> On Fri, Nov 25, 2016 at 4:56 AM, Mikhail Khludnev wrote: >>> >>> Sure. It make sense. Probably, no one added them yet, because if we facet >>> something, it's reasonably that we filter by this field then. >>> >>> On Fri, Nov 25, 2016 at 2:12 AM, Erick Erickson >>> wrote: Am I missing something or do our tests for DocValues=true, indexed=false tests... lack a little something for distributed mode? I'm trying to write a unit test for SOLR-5260 and piggy-backing on DistributedFacetExistsSmallTest. I'm having some problems there, which led me to start looking at other prior art and basically I can't find much. What I want: docValues=true, indexed=false in distributed mode. I found some schemas that have that, notably schema-docValues.xml but I don't find it (or some of the others) in the distributed tests. If we're pushing docValues for grouping, faceting and sorting and claiming that to accomplish that we don't need to have indexed=true we should, of course, have, you know, like, tests. So before I go and create a new test file for this configuration, I wanted to ask if I'm missing any prior art here. Thanks Erick - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org >>> >>> >>> >>> -- >>> Sincerely yours >>> Mikhail Khludnev >> >> > - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
Re: DocValues only faceting tests in distributed mode.
The test doesn't have anything to do with faceting, though. Wasn't sure that you were specifically looking for a faceting test with docValues=true, indexed=false, or just a general one. On Sat, Nov 26, 2016 at 4:48 AM, Ishan Chattopadhyaya < ichattopadhy...@gmail.com> wrote: > Hi Erick, > As part of working on SOLR-5944, I wrote a few tests with docValues=true, > indexed=false, stored=false. You can find it in the latest patch there. The > schema file is called: schema-inplace-updates.xml, and one of the tests > using that schema is: TestInPlaceUpdatesDistrib > Regards, > Ishan > > On Fri, Nov 25, 2016 at 4:56 AM, Mikhail Khludnev wrote: > >> Sure. It make sense. Probably, no one added them yet, because if we facet >> something, it's reasonably that we filter by this field then. >> >> On Fri, Nov 25, 2016 at 2:12 AM, Erick Erickson >> wrote: >> >>> Am I missing something or do our tests for DocValues=true, >>> indexed=false tests... lack a little something for distributed mode? >>> >>> I'm trying to write a unit test for SOLR-5260 and piggy-backing on >>> DistributedFacetExistsSmallTest. I'm having some problems there, which >>> led me to start looking at other prior art and basically I can't find >>> much. >>> >>> What I want: docValues=true, indexed=false in distributed mode. I >>> found some schemas that have that, notably schema-docValues.xml but I >>> don't find it (or some of the others) in the distributed tests. >>> >>> If we're pushing docValues for grouping, faceting and sorting and >>> claiming that to accomplish that we don't need to have indexed=true we >>> should, of course, have, you know, like, tests. >>> >>> So before I go and create a new test file for this configuration, I >>> wanted to ask if I'm missing any prior art here. >>> >>> Thanks >>> Erick >>> >>> - >>> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org >>> For additional commands, e-mail: dev-h...@lucene.apache.org >>> >>> >> >> >> -- >> Sincerely yours >> Mikhail Khludnev >> > >
Re: DocValues only faceting tests in distributed mode.
Hi Erick, As part of working on SOLR-5944, I wrote a few tests with docValues=true, indexed=false, stored=false. You can find it in the latest patch there. The schema file is called: schema-inplace-updates.xml, and one of the tests using that schema is: TestInPlaceUpdatesDistrib Regards, Ishan On Fri, Nov 25, 2016 at 4:56 AM, Mikhail Khludnev wrote: > Sure. It make sense. Probably, no one added them yet, because if we facet > something, it's reasonably that we filter by this field then. > > On Fri, Nov 25, 2016 at 2:12 AM, Erick Erickson > wrote: > >> Am I missing something or do our tests for DocValues=true, >> indexed=false tests... lack a little something for distributed mode? >> >> I'm trying to write a unit test for SOLR-5260 and piggy-backing on >> DistributedFacetExistsSmallTest. I'm having some problems there, which >> led me to start looking at other prior art and basically I can't find >> much. >> >> What I want: docValues=true, indexed=false in distributed mode. I >> found some schemas that have that, notably schema-docValues.xml but I >> don't find it (or some of the others) in the distributed tests. >> >> If we're pushing docValues for grouping, faceting and sorting and >> claiming that to accomplish that we don't need to have indexed=true we >> should, of course, have, you know, like, tests. >> >> So before I go and create a new test file for this configuration, I >> wanted to ask if I'm missing any prior art here. >> >> Thanks >> Erick >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org >> For additional commands, e-mail: dev-h...@lucene.apache.org >> >> > > > -- > Sincerely yours > Mikhail Khludnev >
Re: DocValues only faceting tests in distributed mode.
Sure. It make sense. Probably, no one added them yet, because if we facet something, it's reasonably that we filter by this field then. On Fri, Nov 25, 2016 at 2:12 AM, Erick Erickson wrote: > Am I missing something or do our tests for DocValues=true, > indexed=false tests... lack a little something for distributed mode? > > I'm trying to write a unit test for SOLR-5260 and piggy-backing on > DistributedFacetExistsSmallTest. I'm having some problems there, which > led me to start looking at other prior art and basically I can't find > much. > > What I want: docValues=true, indexed=false in distributed mode. I > found some schemas that have that, notably schema-docValues.xml but I > don't find it (or some of the others) in the distributed tests. > > If we're pushing docValues for grouping, faceting and sorting and > claiming that to accomplish that we don't need to have indexed=true we > should, of course, have, you know, like, tests. > > So before I go and create a new test file for this configuration, I > wanted to ask if I'm missing any prior art here. > > Thanks > Erick > > - > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > > -- Sincerely yours Mikhail Khludnev
DocValues only faceting tests in distributed mode.
Am I missing something or do our tests for DocValues=true, indexed=false tests... lack a little something for distributed mode? I'm trying to write a unit test for SOLR-5260 and piggy-backing on DistributedFacetExistsSmallTest. I'm having some problems there, which led me to start looking at other prior art and basically I can't find much. What I want: docValues=true, indexed=false in distributed mode. I found some schemas that have that, notably schema-docValues.xml but I don't find it (or some of the others) in the distributed tests. If we're pushing docValues for grouping, faceting and sorting and claiming that to accomplish that we don't need to have indexed=true we should, of course, have, you know, like, tests. So before I go and create a new test file for this configuration, I wanted to ask if I'm missing any prior art here. Thanks Erick - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org