Examples of IndexHandlerImpl

com.mysql.clusterj.core.metadata.IndexHandlerImpl

IndexHandlerImpl represents indices defined in the cluster. One instance of this class represents either an ordered index or a unique hash index. So a unique index that results in creating two indices will be represented by two instances of IndexHandlerImpl. The Dictionary access to unique indexes requires the suffix "$unique" to be appended to the name, but the NdbTransaction access requires that the suffix not be used. This class is responsible for mediating the difference.

The simple case is one index=> one field=> one column.

For ClusterJ and JPA, indexes can also support multiple columns, with each column mapped to one field. This pattern is used for compound primary keys. In this case, there is one instance of IndexHandlerImpl for each index, and the columnNames and fields have the same cardinality. This is one index=> multiple (one field=> one column)

For JPA, indexes can also support one field mapped to multiple columns, which is the pattern used for compound foreign keys to represent relationships. In this case, there is a single instance of IndexHandlerImpl for each index. The columnNames lists the columns covered by the index, and there is one field. The field manages an instance of the object id class for the relationship. This is one index=> one field=> multiple columns.

     * @return the array of array identifying the indexes into the IndexHandler
     * list and columns in the IndexHandler corresponding to the field
     */
    public int[][] createIndexHandler( NdbOpenJPADomainFieldHandlerImpl fieldHandler,
            Dictionary dictionary, String indexName) {
        IndexHandlerImpl indexHandler = new IndexHandlerImpl(this, dictionary, indexName, fieldHandler);
        int currentIndex = indexHandlerImpls.size();
        indexHandlerImpls.add(indexHandler);
        int[][] result = new int[1][2];
        result[0][0] = currentIndex;
        result[0][1] = 0;


        // Process indexes for the table. There might not be a field associated with the index.
        // The first entry in indexHandlerImpls is for the mandatory hash primary key,
        // which is not really an index but is treated as an index by query.
        Index primaryIndex = dictionary.getIndex("PRIMARY$KEY", tableName, "PRIMARY");
        IndexHandlerImpl primaryIndexHandler =
            new IndexHandlerImpl(this, dictionary, primaryIndex, primaryKeyColumnNames);
        indexHandlerImpls.add(primaryIndexHandler);


        String[] indexNames = table.getIndexNames();
        for (String indexName: indexNames) {
            // the index alias is the name as known by the user (without the $unique suffix)
            String indexAlias = removeUniqueSuffix(indexName);
            Index index = dictionary.getIndex(indexName, tableName, indexAlias);
            String[] indexColumnNames = index.getColumnNames();
            IndexHandlerImpl imd = new IndexHandlerImpl(this, dictionary, index, indexColumnNames);
            indexHandlerImpls.add(imd);
        }


        // Now iterate the columns in the table, creating a DomainFieldHandler for each column
        for (String columnName: columnNames) {

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.

Examples of IndexHandlerImpl

Examples of com.mysql.clusterj.core.metadata.IndexHandlerImpl

Examples of com.mysql.clusterj.core.metadata.IndexHandlerImpl