Package org.apache.cassandra.index.sasi.plan

Class Operation

  • All Implemented Interfaces:
    com.google.common.collect.PeekingIterator<Token>, java.io.Closeable, java.lang.AutoCloseable, java.util.Iterator<Token>
    public class Operation
extends RangeIterator<java.lang.Long,​Token>

    • Method Detail

      • satisfiedBy

        public boolean satisfiedBy​(Unfiltered currentCluster,
                           Row staticRow,
                           boolean allowMissingColumns)
        Recursive "satisfies" checks based on operation and data from the lower level members using depth-first search and bubbling the results back to the top level caller. Most of the work here is done by localSatisfiedBy(Unfiltered, Row, boolean) see it's comment for details, if there are no local expressions assigned to Operation it will call satisfiedBy(Row) on it's children. Query: first_name = X AND (last_name = Y OR address = XYZ AND street = IL AND city = C) OR (state = 'CA' AND country = 'US') Row: key1: (first_name: X, last_name: Z, address: XYZ, street: IL, city: C, state: NY, country:US) #1 OR / \ #2 (first_name) AND AND (state, country) \ #3 (last_name) OR \ #4 AND (address, street, city) Evaluation of the key1 is top-down depth-first search: --- going down --- Level #1 is evaluated, OR expression has to pull results from it's children which are at level #2 and OR them together, Level #2 AND (state, country) could be be evaluated right away, AND (first_name) refers to it's "right" child from level #3 Level #3 OR (last_name) requests results from level #4 Level #4 AND (address, street, city) does logical AND between it's 3 fields, returns result back to level #3. --- bubbling up --- Level #3 computes OR between AND (address, street, city) result and it's "last_name" expression Level #2 computes AND between "first_name" and result of level #3, AND (state, country) which is already computed Level #1 does OR between results of AND (first_name) and AND (state, country) and returns final result.
        Parameters:
        currentCluster - The row cluster to check.
        staticRow - The static row associated with current cluster.
        allowMissingColumns - allow columns value to be null.
        Returns:
        true if give Row satisfied all of the expressions in the tree, false otherwise.

      • computeNext

        protected Token computeNext()
        Description copied from class: AbstractIterator
        Returns the next element. Note: the implementation must call AbstractIterator.endOfData() when there are no elements left in the iteration. Failure to do so could result in an infinite loop.

        The initial invocation of AbstractIterator.hasNext() or AbstractIterator.next() calls this method, as does the first invocation of hasNext or next following each successful call to next. Once the implementation either invokes endOfData or throws an exception, computeNext is guaranteed to never be called again.

        If this method throws an exception, it will propagate outward to the hasNext or next invocation that invoked this method. Any further attempts to use the iterator will result in an IllegalStateException.

        The implementation of this method may not invoke the hasNext, next, or AbstractIterator.peek() methods on this instance; if it does, an IllegalStateException will result.

        Specified by:
        computeNext in class AbstractIterator<Token>
        Returns:
        the next element if there was one. If endOfData was called during execution, the return value will be ignored.

      • close

        public void close()
           throws java.io.IOException
        Throws:
        java.io.IOException