Examples of IndexMemStore

com.salesforce.hbase.index.covered.data.IndexMemStore
Like the HBase {@link MemStore}, but without all that extra work around maintaining snapshots and sizing (for right now). We still support the concurrent access (in case indexes are built in parallel).
We basically wrap a KeyValueSkipListSet, just like a regular MemStore, except we are:
1. not dealing with
  - space considerations
  - a snapshot set
2. ignoring memstore timestamps in favor of deciding when we want to overwrite keys based on how we obtain them
3. ignoring time range updates (so {@link KeyValueScanner#shouldUseScanner(Scan,SortedSet,long)} isn't supported from {@link #getScanner()}).
We can ignore the memstore timestamps because we know that anything we get from the local region is going to be MVCC visible - so it should just go in. However, we also want overwrite any existing state with our pending write that we are indexing, so that needs to clobber the KVs we get from the HRegion. This got really messy with a regular memstore as each KV from the MemStore frequently has a higher MemStoreTS, but we can't just up the pending KVs' MemStoreTs because a memstore relies on the MVCC readpoint, which generally is less than {@link Long#MAX_VALUE}.
By realizing that we don't need the snapshot or space requirements, we can go much faster than the previous implementation. Further, by being smart about how we manage the KVs, we can drop the extra object creation we were doing to wrap the pending KVs (which we did previously to ensure they sorted before the ones we got from the HRegion). We overwrite {@link KeyValue}s when we add them from external sources {@link #add(KeyValue,boolean)}, but then don't overwrite existing keyvalues when read them from the underlying table (because pending keyvalues should always overwrite current ones) - this logic is all contained in LocalTableState. @see LocalTableState
org.apache.hbase.index.covered.data.IndexMemStore
Like the HBase {@link MemStore}, but without all that extra work around maintaining snapshots and sizing (for right now). We still support the concurrent access (in case indexes are built in parallel).
We basically wrap a KeyValueSkipListSet, just like a regular MemStore, except we are:
1. not dealing with
  - space considerations
  - a snapshot set
2. ignoring memstore timestamps in favor of deciding when we want to overwrite keys based on how we obtain them
3. ignoring time range updates (so {@link KeyValueScanner#shouldUseScanner(Scan,SortedSet,long)} isn't supported from {@link #getScanner()}).
We can ignore the memstore timestamps because we know that anything we get from the local region is going to be MVCC visible - so it should just go in. However, we also want overwrite any existing state with our pending write that we are indexing, so that needs to clobber the KVs we get from the HRegion. This got really messy with a regular memstore as each KV from the MemStore frequently has a higher MemStoreTS, but we can't just up the pending KVs' MemStoreTs because a memstore relies on the MVCC readpoint, which generally is less than {@link Long#MAX_VALUE}.
By realizing that we don't need the snapshot or space requirements, we can go much faster than the previous implementation. Further, by being smart about how we manage the KVs, we can drop the extra object creation we were doing to wrap the pending KVs (which we did previously to ensure they sorted before the ones we got from the HRegion). We overwrite {@link KeyValue}s when we add them from external sources {@link #add(KeyValue,boolean)}, but then don't overwrite existing keyvalues when read them from the underlying table (because pending keyvalues should always overwrite current ones) - this logic is all contained in LocalTableState. @see LocalTableState
org.apache.phoenix.hbase.index.covered.data.IndexMemStore
Like the HBase {@link MemStore}, but without all that extra work around maintaining snapshots and sizing (for right now). We still support the concurrent access (in case indexes are built in parallel).
We basically wrap a KeyValueSkipListSet, just like a regular MemStore, except we are:
1. not dealing with
  - space considerations
  - a snapshot set
2. ignoring memstore timestamps in favor of deciding when we want to overwrite keys based on how we obtain them
3. ignoring time range updates (so {@link KeyValueScanner#shouldUseScanner(Scan,SortedSet,long)} isn't supported from {@link #getScanner()}).
We can ignore the memstore timestamps because we know that anything we get from the local region is going to be MVCC visible - so it should just go in. However, we also want overwrite any existing state with our pending write that we are indexing, so that needs to clobber the KVs we get from the HRegion. This got really messy with a regular memstore as each KV from the MemStore frequently has a higher MemStoreTS, but we can't just up the pending KVs' MemStoreTs because a memstore relies on the MVCC readpoint, which generally is less than {@link Long#MAX_VALUE}.
By realizing that we don't need the snapshot or space requirements, we can go much faster than the previous implementation. Further, by being smart about how we manage the KVs, we can drop the extra object creation we were doing to wrap the pending KVs (which we did previously to ensure they sorted before the ones we got from the HRegion). We overwrite {@link KeyValue}s when we add them from external sources {@link #add(KeyValue,boolean)}, but then don't overwrite existing keyvalues when read them from the underlying table (because pending keyvalues should always overwrite current ones) - this logic is all contained in LocalTableState. @see LocalTableState

Examples of com.salesforce.hbase.index.covered.data.IndexMemStore


  public LocalTableState(RegionCoprocessorEnvironment environment, LocalHBaseState table, Mutation update) {
    this.env = environment;
    this.table = table;
    this.update = update;
    this.memstore = new IndexMemStore();
    this.scannerBuilder = new ScannerBuilder(memstore, update);
    this.columnSet = new CoveredColumns();
  }

View Full Code Here

Examples of org.apache.hbase.index.covered.data.IndexMemStore


  public LocalTableState(RegionCoprocessorEnvironment environment, LocalHBaseState table, Mutation update) {
    this.env = environment;
    this.table = table;
    this.update = update;
    this.memstore = new IndexMemStore();
    this.scannerBuilder = new ScannerBuilder(memstore, update);
    this.columnSet = new CoveredColumns();
  }

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore


  public LocalTableState(RegionCoprocessorEnvironment environment, LocalHBaseState table, Mutation update) {
    this.env = environment;
    this.table = table;
    this.update = update;
    this.memstore = new IndexMemStore();
    this.scannerBuilder = new ScannerBuilder(memstore, update);
    this.columnSet = new CoveredColumns();
  }

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore

  private static final byte[] val = Bytes.toBytes("val");
  private static final byte[] val2 = Bytes.toBytes("val2");


  @Test
  public void testCorrectOverwritting() throws Exception {
    IndexMemStore store = new IndexMemStore(IndexMemStore.COMPARATOR);
    long ts = 10;
    KeyValue kv = new KeyValue(row, family, qual, ts, Type.Put, val);
    kv.setMemstoreTS(2);
    KeyValue kv2 = new KeyValue(row, family, qual, ts, Type.Put, val2);
    kv2.setMemstoreTS(0);
    store.add(kv, true);
    // adding the exact same kv shouldn't change anything stored if not overwritting
    store.add(kv2, false);
    KeyValueScanner scanner = store.getScanner();
    KeyValue first = KeyValue.createFirstOnRow(row);
    scanner.seek(first);
    assertTrue("Overwrote kv when specifically not!", kv == scanner.next());
    scanner.close();


    // now when we overwrite, we should get the newer one
    store.add(kv2, true);
    scanner = store.getScanner();
    scanner.seek(first);
    assertTrue("Didn't overwrite kv when specifically requested!", kv2 == scanner.next());
    scanner.close();
  }

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore

   * {@link Type#DeleteFamily} has a column qualifier specified.
   * @throws Exception
   */
  @Test
  public void testExpectedOrdering() throws Exception {
    IndexMemStore store = new IndexMemStore();
    KeyValue kv = new KeyValue(row, family, qual, 12, Type.Put, val);
    store.add(kv, true);
    KeyValue kv2 = new KeyValue(row, family, qual, 10, Type.Put, val2);
    store.add(kv2, true);
    KeyValue df = new KeyValue(row, family, null, 11, Type.DeleteFamily, null);
    store.add(df, true);
    KeyValue dc = new KeyValue(row, family, qual, 11, Type.DeleteColumn, null);
    store.add(dc, true);
    KeyValue d = new KeyValue(row, family, qual, 12, Type.Delete, null);
    store.add(d, true);


    // null qualifiers should always sort before the non-null cases
    KeyValueScanner scanner = store.getScanner();
    KeyValue first = KeyValue.createFirstOnRow(row);
    assertTrue("Didn't have any data in the scanner", scanner.seek(first));
    assertTrue("Didn't get delete family first (no qualifier == sort first)", df == scanner.next());
    assertTrue("Didn't get point delete before corresponding put", d == scanner.next());
    assertTrue("Didn't get larger ts Put", kv == scanner.next());

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore

  private static final byte[] val = Bytes.toBytes("val");
  private static final byte[] val2 = Bytes.toBytes("val2");


  @Test
  public void testCorrectOverwritting() throws Exception {
    IndexMemStore store = new IndexMemStore(IndexMemStore.COMPARATOR);
    long ts = 10;
    KeyValue kv = new KeyValue(row, family, qual, ts, Type.Put, val);
    kv.setMvccVersion(2);
    KeyValue kv2 = new KeyValue(row, family, qual, ts, Type.Put, val2);
    kv2.setMvccVersion(0);
    store.add(kv, true);
    // adding the exact same kv shouldn't change anything stored if not overwritting
    store.add(kv2, false);
    KeyValueScanner scanner = store.getScanner();
    KeyValue first = KeyValue.createFirstOnRow(row);
    scanner.seek(first);
    assertTrue("Overwrote kv when specifically not!", kv == scanner.next());
    scanner.close();


    // now when we overwrite, we should get the newer one
    store.add(kv2, true);
    scanner = store.getScanner();
    scanner.seek(first);
    assertTrue("Didn't overwrite kv when specifically requested!", kv2 == scanner.next());
    scanner.close();
  }

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore

   * {@link Type#DeleteFamily} has a column qualifier specified.
   * @throws Exception
   */
  @Test
  public void testExpectedOrdering() throws Exception {
    IndexMemStore store = new IndexMemStore();
    KeyValue kv = new KeyValue(row, family, qual, 12, Type.Put, val);
    store.add(kv, true);
    KeyValue kv2 = new KeyValue(row, family, qual, 10, Type.Put, val2);
    store.add(kv2, true);
    KeyValue df = new KeyValue(row, family, null, 11, Type.DeleteFamily, null);
    store.add(df, true);
    KeyValue dc = new KeyValue(row, family, qual, 11, Type.DeleteColumn, null);
    store.add(dc, true);
    KeyValue d = new KeyValue(row, family, qual, 12, Type.Delete, null);
    store.add(d, true);


    // null qualifiers should always sort before the non-null cases
    KeyValueScanner scanner = store.getScanner();
    KeyValue first = KeyValue.createFirstOnRow(row);
    assertTrue("Didn't have any data in the scanner", scanner.seek(first));
    assertTrue("Didn't get delete family first (no qualifier == sort first)", df == scanner.next());
    assertTrue("Didn't get point delete before corresponding put", d == scanner.next());
    assertTrue("Didn't get larger ts Put", kv == scanner.next());

View Full Code Here

Examples of org.apache.phoenix.hbase.index.covered.data.IndexMemStore


  public LocalTableState(RegionCoprocessorEnvironment environment, LocalHBaseState table, Mutation update) {
    this.env = environment;
    this.table = table;
    this.update = update;
    this.memstore = new IndexMemStore();
    this.scannerBuilder = new ScannerBuilder(memstore, update);
    this.columnSet = new CoveredColumns();
  }

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.