Package org.apache.cassandra.db

Class DataRange

  • Direct Known Subclasses:
    DataRange.Paging
    public class DataRange
extends java.lang.Object
    Groups both the range of partitions to query, and the clustering index filter to apply for each partition (for a (partition) range query).

    The main "trick" is that the clustering index filter can only be obtained by providing the partition key on which the filter will be applied. This is necessary when paging range queries, as we might need a different filter for the starting key than for other keys (because the previous page we had queried may have ended in the middle of a partition).

    • Constructor Detail

      • DataRange

        public DataRange​(AbstractBounds<PartitionPosition> range,
                 ClusteringIndexFilter clusteringIndexFilter)
        Creates a DataRange given a range of partition keys and a clustering index filter. The return DataRange will return the same filter for all keys.
        Parameters:
        range - the range over partition keys to use.
        clusteringIndexFilter - the clustering index filter to use.

    • Method Detail

      • allData

        public static DataRange allData​(IPartitioner partitioner)
        Creates a DataRange to query all data (over the whole ring).
        Parameters:
        partitioner - the partitioner in use for the table.
        Returns:
        the newly create DataRange.

      • forTokenRange

        public static DataRange forTokenRange​(Range<Token> tokenRange)
        Creates a DataRange to query all rows over the provided token range.
        Parameters:
        tokenRange - the (partition key) token range to query.
        Returns:
        the newly create DataRange.

      • forKeyRange

        public static DataRange forKeyRange​(Range<PartitionPosition> keyRange)
        Creates a DataRange to query all rows over the provided key range.
        Parameters:
        keyRange - the (partition key) range to query.
        Returns:
        the newly create DataRange.

      • allData

        public static DataRange allData​(IPartitioner partitioner,
                                ClusteringIndexFilter filter)
        Creates a DataRange to query all partitions of the ring using the provided clustering index filter.
        Parameters:
        partitioner - the partitioner in use for the table queried.
        filter - the clustering index filter to use.
        Returns:
        the newly create DataRange.

      • keyRange

        public AbstractBounds<PartitionPosition> keyRange()
        The range of partition key queried by this DataRange.
        Returns:
        the range of partition key queried by this DataRange.

      • startKey

        public PartitionPosition startKey()
        The start of the partition key range queried by this DataRange.
        Returns:
        the start of the partition key range queried by this DataRange.

      • stopKey

        public PartitionPosition stopKey()
        The end of the partition key range queried by this DataRange.
        Returns:
        the end of the partition key range queried by this DataRange.

      • startAsByteComparable

        public ByteComparable startAsByteComparable()
        The start of the partition key range queried by this DataRange.
        Returns:
        the start of the partition key range expressed as a ByteComparable.

      • stopAsByteComparable

        public ByteComparable stopAsByteComparable()
        The end of the partition key range queried by this DataRange.
        Returns:
        the end of the partition key range expressed as a ByteComparable.

      • isNamesQuery

        public boolean isNamesQuery()
        Whether the underlying clustering index filter is a names filter or not.
        Returns:
        Whether the underlying clustering index filter is a names filter or not.

      • isPaging

        public boolean isPaging()
        Whether the data range is for a paged request or not.
        Returns:
        true if for paging, false otherwise

      • isWrapAround

        public boolean isWrapAround()
        Whether the range queried by this DataRange actually wraps around.
        Returns:
        whether the range queried by this DataRange actually wraps around.

      • contains

        public boolean contains​(PartitionPosition pos)
        Whether the provided ring position is covered by this DataRange.
        Returns:
        whether the provided ring position is covered by this DataRange.

      • isUnrestricted

        public boolean isUnrestricted​(TableMetadata metadata)
        Whether this DataRange queries everything (has no restriction neither on the partition queried, nor within the queried partition).
        Returns:
        Whether this DataRange queries everything.

      • selectsAllPartition

        public boolean selectsAllPartition()

      • isReversed

        public boolean isReversed()
        Whether the underlying ClusteringIndexFilter is reversed or not.
        Returns:
        whether the underlying ClusteringIndexFilter is reversed or not.

      • clusteringIndexFilter

        public ClusteringIndexFilter clusteringIndexFilter​(DecoratedKey key)
        The clustering index filter to use for the provided key.

        This may or may not be the same filter for all keys (that is, paging range use a different filter for their start key).

        Parameters:
        key - the partition key for which we want the clustering index filter.
        Returns:
        the clustering filter to use for key.

      • forPaging

        public DataRange forPaging​(AbstractBounds<PartitionPosition> range,
                           ClusteringComparator comparator,
                           Clustering<?> lastReturned,
                           boolean inclusive)
        Returns a new DataRange for use when paging this range.
        Parameters:
        range - the range of partition keys to query.
        comparator - the comparator for the table queried.
        lastReturned - the clustering for the last result returned by the previous page, i.e. the result we want to start our new page from. This last returned must correspond to left bound of range (in other words, range.left must be the partition key for that lastReturned result).
        inclusive - whether or not we want to include the lastReturned in the newly returned page of results.
        Returns:
        a new DataRange suitable for paging this range given the lastRetuned result of the previous page.

      • forSubRange

        public DataRange forSubRange​(AbstractBounds<PartitionPosition> range)
        Returns a new DataRange equivalent to this one but restricted to the provided sub-range.
        Parameters:
        range - the sub-range to use for the newly returned data range. Note that assumes that range is a proper sub-range of the initial range but doesn't validate it. You should make sure to only provided sub-ranges however or this might throw off the paging case (see Paging.forSubRange()).
        Returns:
        a new DataRange using range as partition key range and the clustering index filter filter from this.

      • toString

        public java.lang.String toString​(TableMetadata metadata)

      • appendKeyString

        public static void appendKeyString​(java.lang.StringBuilder sb,
                                   AbstractType<?> type,
                                   java.nio.ByteBuffer key)