garrrruii/rtreeproject

Rtree building and query processing

Rtree building and query processing

2D-R-tree code

2020-11-13

• 2D-Rtree of point object for blockchain => no need deletion
• count logical IO of tree-building and range/knn query processing => node access count

NOW:

• Point with integer coordinates for object, 2D-rectangle(mbr)---node
• INSERT: insert obj to node with 1)smallest enlarged area 2)smallest area
• DELETE: NO DELETION.
• SPLIT:
  • divide node in half and half
  • pick two seeds by choosing two nodes making largest wasted space
  • distribute child by choosing what and where to insert with 1)smallest enlarged area 2)smallest area
• QUERYS: range, knn
• COUNT IO: ONLY for this project. Count it logically. Assume that
  1. We have nodes in DISK from lv3 i.e. ONLY root~lv2 nodes are in MEMORY
  2. Parent node can see children's coord info

PLAN:

• Change code using 'Class'
• Change it to apply to rectangular objects or real number coordinates
• If it's possible: visualization using python or else (something helpful to understand Rtree mechanism)
• Add deletion...?
• Add other queries? ex) window, rnn
• Else, working on the project personally (Regardless of THIS project)