[ASTERIXDB-2389] Integrate new ESRI-based spatial support with R-Trees

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.9.4
• Fix Version/s: 0.9.8
• Component/s: COMP - Compiler, DOC - Documentation, IDX - Indexes, MTD - Metadata, SQL - Translator SQL++, STO - Storage, TYPE - Data Model
• Assignee: Ahmed Eldawy
• Reporter: Michael J. Carey
• Created: 19/May/18 15:18
• Updated: 17/Feb/22 01:56

Description

This is the starting point in the codebase to add geometry type to RTree index.
 
Index creation:
QueryTranslator.java

• handleCreateIndexStatement()
    - validateIndexKeyFields() : checks the key field type
       - ValidateUtil.validateKeyFields()
                  case RTREE:
                      switch (fieldType.getTypeTag()) {                         case POINT:                         case LINE:                         case RECTANGLE:                         case CIRCLE:                         case POLYGON:                         case UNION:                             break;                         default:                             throw new AsterixException("The field \"" + fieldName + "\" which is of type "                                     + fieldType.getTypeTag() + " cannot be indexed using the RTree index.");                     }

                      break;
   
   - doCreateIndex
     - IndexUtil.buildSecondaryIndexCreationJobSpec()
       - SecondaryIndexOperationsHelper.createIndexOperationsHelper()
         -      case RTREE:
                  indexOperationsHelper =
                          new SecondaryRTreeOperationsHelper(dataset, index, physOptConf, metadataProvider);
                  break;
     - runJob(hcc, spec, jobFlags);
   
   
  For picking an index, the RTree relevant class is RTreeAccessMethod. 
     
  Index pick during a query:
  IntroduceSelectAccessMethodRule.java: select access transformation
  IntroduceJoinAccessMethodRule.java: join access transformation
  RTreeAccessMethod: specific to RTree
   - FUNC_IDENTIFIERS has a list of supported functions.
   - createIndexSearchPlan: creates an MBR for the search predicate - currently for "spatial_intersect" function only
   - So, for each function that you want to support, you need to generate a suitable MBR that can generate a candidate set
     of the actual answer.