Given a query, Query Executor first parses the
query and classifies it as either non-join query or join query.
- For a non-join query, which involves only a single relation, if
the query has a filter condition (where clause), Query Executor
looks for data indices to process query. Otherwise, Query
Executor searches table indices and column indices for query
processing. Based on retrieved indices, Query Executor then
sends the query to owner peers containing the data.
- For an equal join query, we employ the classical
bloom join algorithm to process it.
|
