Apparatus and method for incremental physical data clustering

A data storage and retrieval system wherein tree-structured data is arranged in nodes and is stored and retrieved in pages, each page comprising a plurality of nodes, the system implemented using an electronic processor, the system comprising:

• a monitoring means for monitoring ongoing data retrieval of said tree data to find retrieval patterns of nodes which are retrieved in temporal proximity and to identify changes in said retrieval patterns over time;
• and
• a node rearranging means for periodically rearranging the data nodes among said pages dynamically during usage of the data to reflect said changes in data usage so that nodes more often retrieved together in temporal proximity are migrated to cluster on shared pages and nodes more often required separately are migrated towards separate pages, thereby to keep small an overall number of page accesses during data retrieval despite dynamic changes in patterns of data retrieval, said periodically rearranging being triggered according to pages.