Examples of MonotonicBlockPackedWriter

org.apache.lucene.util.packed.MonotonicBlockPackedWriter
A writer for large monotonically increasing sequences of positive longs.
The sequence is divided into fixed-size blocks and for each block, values are modeled after a linear function f: x → A × x + B. The block encodes deltas from the expected values computed from this function using as few bits as possible.
Format:
- <BLock>^BlockCount
- BlockCount: ⌈ ValueCount / BlockSize ⌉
- Block: <Header, (Ints)>
- Header: <B, A, BitsPerValue>
- B: the B from f: x → A × x + B using a {@link BitUtil#zigZagEncode(long) zig-zag encoded}{@link DataOutput#writeVLong(long) vLong}
- A: the A from f: x → A × x + B encoded using {@link Float#floatToIntBits(float)} on{@link DataOutput#writeInt(int) 4 bytes}
- BitsPerValue: a {@link DataOutput#writeVInt(int) variable-length int}
- Ints: if BitsPerValue is 0, then there is nothing to read and all values perfectly match the result of the function. Otherwise, these are the {@link PackedInts packed} deltas from the expected value(computed from the function) using exaclty BitsPerValue bits per value.
@see MonotonicBlockPackedReader @lucene.internal

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    // otherwise, we need to record the length fields...
    if (minLength != maxLength) {
      meta.writeVInt(PackedInts.VERSION_CURRENT);
      meta.writeVInt(BLOCK_SIZE);


      final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
      long addr = 0;
      for (BytesRef v : values) {
        if (v != null) {
          addr += v.length;
        }
        writer.add(addr);
      }
      writer.finish();
    }
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    // otherwise, we need to record the length fields...
    if (minLength != maxLength) {
      meta.writeVInt(PackedInts.VERSION_CURRENT);
      meta.writeVInt(BLOCK_SIZE);


      final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
      long addr = 0;
      for (BytesRef v : values) {
        if (v != null) {
          addr += v.length;
        }
        writer.add(addr);
      }
      writer.finish();
    }
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    if (minLength != maxLength) {
      meta.writeLong(data.getFilePointer());
      meta.writeVInt(PackedInts.VERSION_CURRENT);
      meta.writeVInt(BLOCK_SIZE);


      final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
      long addr = 0;
      for (BytesRef v : values) {
        if (v != null) {
          addr += v.length;
        }
        writer.add(addr);
      }
      writer.finish();
    }
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

      // now write the bytes: sharing prefixes within a block
      final long startFP = data.getFilePointer();
      // currently, we have to store the delta from expected for every 1/nth term
      // we could avoid this, but its not much and less overall RAM than the previous approach!
      RAMOutputStream addressBuffer = new RAMOutputStream();
      MonotonicBlockPackedWriter termAddresses = new MonotonicBlockPackedWriter(addressBuffer, BLOCK_SIZE);
      BytesRef lastTerm = new BytesRef();
      long count = 0;
      for (BytesRef v : values) {
        if (count % ADDRESS_INTERVAL == 0) {
          termAddresses.add(data.getFilePointer() - startFP);
          // force the first term in a block to be abs-encoded
          lastTerm.length = 0;
        }
        
        // prefix-code
        int sharedPrefix = StringHelper.bytesDifference(lastTerm, v);
        data.writeVInt(sharedPrefix);
        data.writeVInt(v.length - sharedPrefix);
        data.writeBytes(v.bytes, v.offset + sharedPrefix, v.length - sharedPrefix);
        lastTerm.copyBytes(v);
        count++;
      }
      final long indexStartFP = data.getFilePointer();
      // write addresses of indexed terms
      termAddresses.finish();
      addressBuffer.writeTo(data);
      addressBuffer = null;
      termAddresses = null;
      meta.writeVInt(minLength);
      meta.writeVInt(maxLength);

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    meta.writeVInt(PackedInts.VERSION_CURRENT);
    meta.writeLong(data.getFilePointer());
    meta.writeVLong(maxDoc);
    meta.writeVInt(BLOCK_SIZE);


    final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
    long addr = 0;
    for (Number v : docToOrdCount) {
      addr += v.longValue();
      writer.add(addr);
    }
    writer.finish();
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

      // now write the bytes: sharing prefixes within a block
      final long startFP = data.getFilePointer();
      // currently, we have to store the delta from expected for every 1/nth term
      // we could avoid this, but its not much and less overall RAM than the previous approach!
      RAMOutputStream addressBuffer = new RAMOutputStream();
      MonotonicBlockPackedWriter termAddresses = new MonotonicBlockPackedWriter(addressBuffer, BLOCK_SIZE);
      BytesRef lastTerm = new BytesRef();
      long count = 0;
      for (BytesRef v : values) {
        if (count % ADDRESS_INTERVAL == 0) {
          termAddresses.add(data.getFilePointer() - startFP);
          // force the first term in a block to be abs-encoded
          lastTerm.length = 0;
        }
        
        // prefix-code
        int sharedPrefix = StringHelper.bytesDifference(lastTerm, v);
        data.writeVInt(sharedPrefix);
        data.writeVInt(v.length - sharedPrefix);
        data.writeBytes(v.bytes, v.offset + sharedPrefix, v.length - sharedPrefix);
        lastTerm.copyBytes(v);
        count++;
      }
      final long indexStartFP = data.getFilePointer();
      // write addresses of indexed terms
      termAddresses.finish();
      addressBuffer.writeTo(data);
      addressBuffer = null;
      termAddresses = null;
      meta.writeVInt(minLength);
      meta.writeVInt(maxLength);

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    meta.writeVInt(PackedInts.VERSION_CURRENT);
    meta.writeLong(data.getFilePointer());
    meta.writeVLong(maxDoc);
    meta.writeVInt(BLOCK_SIZE);


    final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
    long addr = 0;
    for (Number v : docToOrdCount) {
      addr += v.longValue();
      writer.add(addr);
    }
    writer.finish();
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    if (minLength != maxLength) {
      meta.writeLong(data.getFilePointer());
      meta.writeVInt(PackedInts.VERSION_CURRENT);
      meta.writeVInt(BLOCK_SIZE);


      final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
      long addr = 0;
      writer.add(addr);
      for (BytesRef v : values) {
        if (v != null) {
          addr += v.length;
        }
        writer.add(addr);
      }
      writer.finish();
    }
  }

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

      // now write the bytes: sharing prefixes within a block
      final long startFP = data.getFilePointer();
      // currently, we have to store the delta from expected for every 1/nth term
      // we could avoid this, but its not much and less overall RAM than the previous approach!
      RAMOutputStream addressBuffer = new RAMOutputStream();
      MonotonicBlockPackedWriter termAddresses = new MonotonicBlockPackedWriter(addressBuffer, BLOCK_SIZE);
      BytesRef lastTerm = new BytesRef();
      long count = 0;
      for (BytesRef v : values) {
        if (count % ADDRESS_INTERVAL == 0) {
          termAddresses.add(data.getFilePointer() - startFP);
          // force the first term in a block to be abs-encoded
          lastTerm.length = 0;
        }
        
        // prefix-code
        int sharedPrefix = StringHelper.bytesDifference(lastTerm, v);
        data.writeVInt(sharedPrefix);
        data.writeVInt(v.length - sharedPrefix);
        data.writeBytes(v.bytes, v.offset + sharedPrefix, v.length - sharedPrefix);
        lastTerm.copyBytes(v);
        count++;
      }
      final long indexStartFP = data.getFilePointer();
      // write addresses of indexed terms
      termAddresses.finish();
      addressBuffer.writeTo(data);
      addressBuffer = null;
      termAddresses = null;
      meta.writeVInt(minLength);
      meta.writeVInt(maxLength);

View Full Code Here

Examples of org.apache.lucene.util.packed.MonotonicBlockPackedWriter

    meta.writeLong(data.getFilePointer());
    meta.writeVLong(maxDoc);
    meta.writeVInt(PackedInts.VERSION_CURRENT);
    meta.writeVInt(BLOCK_SIZE);


    final MonotonicBlockPackedWriter writer = new MonotonicBlockPackedWriter(data, BLOCK_SIZE);
    long addr = 0;
    writer.add(addr);
    for (Number v : values) {
      addr += v.longValue();
      writer.add(addr);
    }
    writer.finish();
    meta.writeLong(data.getFilePointer());
  }

View Full Code Here

0 1 2

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.