Vazquez, Sorry I don't have an answer but I'd love to know what you need this for :-)
I think the logic is going to have to bleed into your search app. In short copy field and your app knows which to search in. lee c On 30 April 2012 20:41, Erick Erickson <erickerick...@gmail.com> wrote: > OK, I took another look at what you were trying to > accomplish and, I find the use-case kind of hard to > figure out, but that's my problem <G>. > > But it is true that there's really no good way to _change_ the > way the field is analyzed in Solr. Of course since Solr is > built on Lucene, you could to a lot of work and bypass the > analysis chain, but that seems like more work than this is > really worth. So using multiple fields seems like the way > to go. > > This is pretty unusual. You have a field that you can >> search and display the first value >> search but not display values 2-99 >> display but not search values 100-* > > nobody's thought about how to specify this in a schema file, > and it appears to be an unusual enough use case that > nobody has proposed anything like that that I know of. > > Best > Erick > > On Mon, Apr 30, 2012 at 1:11 PM, Vazquez, Maria (STM) > <maria.vazq...@dexone.com> wrote: >> Thanks Erick. >> I'm not concerned about the logic, all I want to achieve is sometimes >> storing/indexing a multi-valued field and sometimes not (same field with >> same name) based on some logic. In a transformer I cannot change the >> schema dynamically to do that, not that I know of at least. >> So if I define that field in the schema.xml I guess I'm stuck with the >> store/indexed true/false and cannot change that at indexing time. >> In Lucene you can just do what I had in my original email and the field >> will be sometimes stored/indexed and sometimes not without a problem. >> >> Does that make sense? >> Thanks, >> Maria >> >> >> >> Maria Vazquez | Manager Software Engineering | Dex One >> Phone: 310.586.4157 >> >> www.DexOne.com <http://www.dexone.com/> | www.DexKnows.com >> <http://www.dexknows.com/> >> <http://www.dexone.com/> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> On 4/30/12 6:08 AM, "Erick Erickson" <erickerick...@gmail.com> wrote: >> >>>Your idea of using a Transformer will work just fine, you have a lot more >>>flexibility in a custom Transformer, see: >>>http://wiki.apache.org/solr/DIHCustomTransformer >>> >>>You could also write a custom update handler that examined the >>>document on the server side and implemented your logic, or even >>>just add an element into an "update chain". >>> >>>None of these is very hard, but they can be a bit intimidating >>>when stating from scratch. >>> >>>Best >>>Erick >>> >>>On Sun, Apr 29, 2012 at 12:53 PM, Vazquez, Maria (STM) >>><maria.vazq...@dexone.com> wrote: >>>> Thanks for your response. >>>> That's what I need, changes at indexing time. Dynamic fields are not >>>>what I need because the field name is the same, I just need to change if >>>>they indexed/stored based on some logic. >>>> It's so easily achieved with the Lucene API, I was sure there was a way >>>>to do the same in Solr. >>>> >>>> >>>> On Apr 28, 2012, at 22:34, "Jeevanandam" <je...@myjeeva.com> wrote: >>>> >>>>> Maria, >>>>> >>>>> thanks for detailed explanation. >>>>> as per schema.xml; stored or indexed should be defined at design-time. >>>>>Per my understanding defining at runtime is not feasible. >>>>> BTW, you can have multiValued="true" attribute for dynamic fields too. >>>>> >>>>> - Jeevanandam >>>>> >>>>> On 29-04-2012 2:06 am, Vazquez, Maria (STM) wrote: >>>>>> Thanks Jeevanandam. >>>>>> That still doesn't have the same behavior as Lucene since multiple >>>>>> fields with different names have to be created. >>>>>> What I want is this exactly (multi-value field) >>>>>> >>>>>> document.add(new Field("geoids", geoId, Field.Store.YES, >>>>>> Field.Index.NOT_ANALYZED_NO_NORMS)); >>>>>> >>>>>> document.add(new Field("geoids", geoId, Field.Store.NO, >>>>>> Field.Index.NOT_ANALYZED_NO_NORMS)); >>>>>> >>>>>> In Lucene I can save geoids first as stored and in the next line as >>>>>> not stored and it will do exactly that. I want to duplicate this >>>>>> behavior in Solr but I can't do it having only one field in the schema >>>>>> called geoids that I an manipulate at inde time whether to store or >>>>>> not depending on a condition. >>>>>> >>>>>> Thanks again for the help, hope this explanation makes it more clear >>>>>> in what I'm trying to do. >>>>>> >>>>>> Maria >>>>>> >>>>>> On Apr 28, 2012, at 11:49 AM, "Jeevanandam" >>>>>> <je...@myjeeva.com<mailto:je...@myjeeva.com>> wrote: >>>>>> >>>>>> Maria, >>>>>> >>>>>> For your need please define unique pattern using dynamic field in >>>>>>schema.xml >>>>>> >>>>>> Please have a look >>>>>>http://wiki.apache.org/solr/SchemaXml#Dynamic_fields >>>>>> >>>>>> Hope that helps! >>>>>> >>>>>> -Jeevanandam >>>>>> >>>>>> Technology keeps you connected! >>>>>> >>>>>> On Apr 28, 2012, at 10:33 PM, "Vazquez, Maria (STM)" >>>>>> <maria.vazq...@dexone.com<mailto:maria.vazq...@dexone.com>> wrote: >>>>>> >>>>>> I can call a script for the logic part but what I want to figure out >>>>>> is how to save the same field sometimes as stored and indexed, >>>>>> sometimes as stored not indexed, etc. From a transformer or a script I >>>>>> didn't see anything where I can modify that at indexing time. >>>>>> Thanks a lot, >>>>>> Maria >>>>>> >>>>>> >>>>>> On Apr 27, 2012, at 18:38, "Bill Bell" >>>>>> <billnb...@gmail.com<mailto:billnb...@gmail.com>> wrote: >>>>>> >>>>>> Yes you can. Just use a script that is called for each row. >>>>>> >>>>>> Bill Bell >>>>>> Sent from mobile >>>>>> >>>>>> >>>>>> On Apr 27, 2012, at 6:38 PM, "Vazquez, Maria (STM)" >>>>>> <maria.vazq...@dexone.com<mailto:maria.vazq...@dexone.com>> wrote: >>>>>> >>>>>> Hi, >>>>>> I'm migrating a project from Lucene 2.9 to Solr 3.4. >>>>>> There is a special case in the code that indexes the same field in >>>>>> two different ways, which is completely legal in Lucene directly but I >>>>>> don't know how to duplicate this same behavior in Solr: >>>>>> >>>>>> if (isFirstGeo) { >>>>>> document.add(new Field("geoids", geoId, Field.Store.YES, >>>>>> Field.Index.NOT_ANALYZED_NO_NORMS)); >>>>>> isFirstGeo = false; >>>>>> } else { >>>>>> if (countProducts < 100) >>>>>> document.add(new Field("geoids", geoId, Field.Store.NO, >>>>>> Field.Index.NOT_ANALYZED_NO_NORMS)); >>>>>> else >>>>>> document.add(new Field("geoids", geoId, Field.Store.YES, >>>>>> Field.Index.NO)); >>>>>> } >>>>>> >>>>>> Is there any way to do this in Solr in a Tranformer? I'm using the >>>>>> DIH to index and I can't see a way to do this other than having three >>>>>> fields in the schema like geoids_store_index, geoids_nostore_index, >>>>>> and geoids_store_noindex. >>>>>> >>>>>> Thanks a lot in advance. >>>>>> Maria >>>>> >>
