Re: [geomesa-users] Full table scan on attribute query when attribute is

[Emilio Lahr-Vivaz <elahrvivaz@xxxxxxxx>]

I believe we hit something like this before - the EQUALS predicate 
you're using is not what you think it is. Can you try using 'ff.equals' 
instead of 'ff.equal'? The later is org.opengis.filter.spatial.Equals, 
which only operates on geometries, the former is 
org.opengis.filter.PropertyIsEqualTo, which is what you want.

Thanks,

On 07/13/2017 05:29 AM, David Boyd wrote:

Emilio
> All:
>
>    Just got a strange issue.   I have a fairly large featuretype (>8M).
> In my code I am looking to delete a given feature using this code:
>
> >         FilterFactory2 ff = CommonFactoryFinder.getFilterFactory2();
> >         SimpleFeatureStore featureStore = (SimpleFeatureStore) 
> > store.getFeatureSource(LINKAGE_FEATURE_NAME);
> >         Filter filter 
> > =ff.equal(ff.property(LINKAGE_ENTITY1_KEY_COLUMN_NAME),ff.literal(entity.getKey()));
> >
> >         featureStore.removeFeatures(filter);
> My describe schema on the type gives:
>
> > [root@bdpgeoserver geomesa-accumulo_2.11-1.3.1]# bin/geomesa 
> > describe-schema -z bdpnode4 -i rdk -u root -c CoalesceSearch -f Linkages
> > Password (mask enabled)>
> > INFO  Describing attributes of feature 'Linkages'
> > objectKey      | String
> > entity1Key     | String  (Attribute indexed - join)
> > entity1Name    | String
> > entity1Source  | String
> > entity1Version | String
> > entity2Key     | String  (Attribute indexed - join)
> > entity2Name    | String
> > entity2Source  | String
> > entity2Version | String
> > lastModified   | Date    (Spatio-temporally indexed)
> > label          | String  (Attribute indexed - join)
> > linkType       | String  (Attribute indexed - join)
> > theWorld       | Polygon (Spatially indexed)
> >
> > User data:
> >   geomesa.index.dtg            | lastModified
> >   geomesa.indices              | records:2:3,attr:4:3
> >   geomesa.table.sharing        | true
> >   geomesa.table.sharing.prefix |
> But it get this log message for each delete:
>
> 2017-07-13 09:27:32,976 |  WARN | [main] | (IdIndex.scala:77) - 
> Running full table scan for schema Linkages with filter 
> EQUALS(entity1Key, 'd01bb4ac-86ec-4035-86f2-1da765d1e3e9')
> 2017-07-13 09:27:32,977 |  WARN | [main] | (IdIndex.scala:87) - 
> Running full table scan for schema Linkages with filter 
> EQUALS(entity1Key, 'd01bb4ac-86ec-4035-86f2-1da765d1e3e9')
>
> Needless to say each one of these deletes is taking 20-30 seconds.
>
> Thoughts?