Examples of SpillableGroupByCache

com.salesforce.phoenix.cache.aggcache.SpillableGroupByCache
f.uni-tuebingen.de/files/teaching/ws1011/db2/db2-hash-indexes.pdf . For this, each SpillFile keeps a directory of pointers to Integer.MAX_VALUE 4K pages in memory, which allows each directory to address more pages than a single memory mapped temp file could theoretically store. In case directory doubling, requests a page index that exceeds the limits of the initial temp file limits, the implementation dynamically allocates additional temp files to the SpillFile. The directory starts with a global depth of 1 and therefore a directory size of 2 buckets. Only during bucket split and directory doubling more than one page is temporarily kept in memory until all elements have been redistributed. The current implementation conducts bucket splits as long as an element does not fit onto a page. No overflow chain is created, which might be an alternative. For get requests, each directory entry maintains a bloomFilter to prevent page-in operations in case an element has never been spilled before. The deserialization is only triggered when a key a loaded back into the LRU cache. The aggregators are returned from the LRU cache and the next value is computed. In case the key is not found on any page, the Loader create new aggregators for it.
org.apache.phoenix.cache.aggcache.SpillableGroupByCache
f.uni-tuebingen.de/files/teaching/ws1011/db2/db2-hash-indexes.pdf . For this, each SpillFile keeps a directory of pointers to Integer.MAX_VALUE 4K pages in memory, which allows each directory to address more pages than a single memory mapped temp file could theoretically store. In case directory doubling, requests a page index that exceeds the limits of the initial temp file limits, the implementation dynamically allocates additional temp files to the SpillFile. The directory starts with a global depth of 1 and therefore a directory size of 2 buckets. Only during bucket split and directory doubling more than one page is temporarily kept in memory until all elements have been redistributed. The current implementation conducts bucket splits as long as an element does not fit onto a page. No overflow chain is created, which might be an alternative. For get requests, each directory entry maintains a bloomFilter to prevent page-in operations in case an element has never been spilled before. The deserialization is only triggered when a key a loaded back into the LRU cache. The aggregators are returned from the LRU cache and the next value is computed. In case the key is not found on any page, the Loader create new aggregators for it.

Examples of com.salesforce.phoenix.cache.aggcache.SpillableGroupByCache

        GroupByCache newCache(RegionCoprocessorEnvironment env, ImmutableBytesWritable tenantId, ServerAggregators aggregators, int estDistVals) {
            Configuration conf = env.getConfiguration();
            boolean spillableEnabled =
                    conf.getBoolean(GROUPBY_SPILLABLE_ATTRIB, DEFAULT_GROUPBY_SPILLABLE);
            if (spillableEnabled) {
                return new SpillableGroupByCache(env, tenantId, aggregators, estDistVals);
            } 
            
            return new InMemoryGroupByCache(env, tenantId, aggregators, estDistVals);
        }

View Full Code Here

Examples of org.apache.phoenix.cache.aggcache.SpillableGroupByCache

        GroupByCache newCache(RegionCoprocessorEnvironment env, ImmutableBytesWritable tenantId, ServerAggregators aggregators, int estDistVals) {
            Configuration conf = env.getConfiguration();
            boolean spillableEnabled =
                    conf.getBoolean(GROUPBY_SPILLABLE_ATTRIB, DEFAULT_GROUPBY_SPILLABLE);
            if (spillableEnabled) {
                return new SpillableGroupByCache(env, tenantId, aggregators, estDistVals);
            } 
            
            return new InMemoryGroupByCache(env, tenantId, aggregators, estDistVals);
        }

View Full Code Here

Examples of org.apache.phoenix.cache.aggcache.SpillableGroupByCache

        GroupByCache newCache(RegionCoprocessorEnvironment env, ImmutableBytesWritable tenantId, ServerAggregators aggregators, int estDistVals) {
            Configuration conf = env.getConfiguration();
            boolean spillableEnabled =
                    conf.getBoolean(GROUPBY_SPILLABLE_ATTRIB, DEFAULT_GROUPBY_SPILLABLE);
            if (spillableEnabled) {
                return new SpillableGroupByCache(env, tenantId, aggregators, estDistVals);
            } 
            
            return new InMemoryGroupByCache(env, tenantId, aggregators, estDistVals);
        }

View Full Code Here

Examples of org.apache.phoenix.cache.aggcache.SpillableGroupByCache

        GroupByCache newCache(RegionCoprocessorEnvironment env, ImmutableBytesWritable tenantId, ServerAggregators aggregators, int estDistVals) {
            Configuration conf = env.getConfiguration();
            boolean spillableEnabled =
                    conf.getBoolean(GROUPBY_SPILLABLE_ATTRIB, DEFAULT_GROUPBY_SPILLABLE);
            if (spillableEnabled) {
                return new SpillableGroupByCache(env, tenantId, aggregators, estDistVals);
            } 
            
            return new InMemoryGroupByCache(env, tenantId, aggregators, estDistVals);
        }

View Full Code Here

Examples of org.apache.phoenix.cache.aggcache.SpillableGroupByCache

        GroupByCache newCache(RegionCoprocessorEnvironment env, ImmutableBytesWritable tenantId, ServerAggregators aggregators, int estDistVals) {
            Configuration conf = env.getConfiguration();
            boolean spillableEnabled =
                    conf.getBoolean(GROUPBY_SPILLABLE_ATTRIB, DEFAULT_GROUPBY_SPILLABLE);
            if (spillableEnabled) {
                return new SpillableGroupByCache(env, tenantId, aggregators, estDistVals);
            } 
            
            return new InMemoryGroupByCache(env, tenantId, aggregators, estDistVals);
        }

View Full Code Here

TOP

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.