RussellSpitzer commented on a change in pull request #3966:
URL: https://github.com/apache/iceberg/pull/3966#discussion_r802098213



##########
File path: core/src/main/java/org/apache/iceberg/util/ZOrderByteUtils.java
##########
@@ -0,0 +1,187 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+package org.apache.iceberg.util;
+
+import java.nio.ByteBuffer;
+import java.nio.charset.StandardCharsets;
+import java.util.Arrays;
+
+/**
+ * Within Z-Ordering the byte representations of objects being compared must 
be ordered,
+ * this requires several types to be transformed when converted to bytes. The 
goal is to
+ * map object's whose byte representation are not lexicographically ordered 
into representations
+ * that are lexicographically ordered.
+ * Most of these techniques are derived from
+ * 
https://aws.amazon.com/blogs/database/z-order-indexing-for-multifaceted-queries-in-amazon-dynamodb-part-2/
+ *
+ * Some implementation is taken from
+ * 
https://github.com/apache/hbase/blob/master/hbase-common/src/main/java/org/apache/hadoop/hbase/util/OrderedBytes.java
+ */
+public class ZOrderByteUtils {
+
+  private ZOrderByteUtils() {
+
+  }
+
+  /**
+   * Signed ints do not have their bytes in magnitude order because of the 
sign bit.
+   * To fix this, flip the sign bit so that all negatives are ordered before 
positives. This essentially
+   * shifts the 0 value so that we don't break our ordering when we cross the 
new 0 value.
+   */
+  public static byte[] intToOrderedBytes(int val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Integer.BYTES);
+    bytes.putInt(val ^ 0x80000000);
+    return bytes.array();
+  }
+
+  /**
+   * Signed longs are treated the same as the signed ints in {@link 
#intToOrderedBytes(int, ByteBuffer)}
+   */
+  public static byte[] longToOrderedBytes(long val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Long.BYTES);
+    bytes.putLong(val ^ 0x8000000000000000L);
+    return bytes.array();
+  }
+
+  /**
+   * Signed shorts are treated the same as the signed ints in {@link 
#intToOrderedBytes(int, ByteBuffer)}
+   */
+  public static byte[] shortToOrderedBytes(short val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Short.BYTES);
+    bytes.putShort((short) (val ^ (0x8000)));
+    return bytes.array();
+  }
+
+  /**
+   * Signed tiny ints are treated the same as the signed ints in {@link 
#intToOrderedBytes(int, ByteBuffer)}
+   */
+  public static byte[] tinyintToOrderedBytes(byte val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Byte.BYTES);
+    bytes.put((byte) (val ^ (0x80)));
+    return bytes.array();
+  }
+
+  /**
+   * IEEE 754 :
+   * “If two floating-point numbers in the same format are ordered (say, x 
{@literal <} y),
+   * they are ordered the same way when their bits are reinterpreted as 
sign-magnitude integers.”
+   *
+   * Which means floats can be treated as sign magnitude integers which can 
then be converted into lexicographically
+   * comparable bytes
+   */
+  public static byte[] floatToOrderedBytes(float val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Float.BYTES);
+    int ival = Float.floatToIntBits(val);
+    ival ^= ((ival >> (Integer.SIZE - 1)) | Integer.MIN_VALUE);
+    bytes.putInt(ival);
+    return bytes.array();
+  }
+
+  /**
+   * Doubles are treated the same as floats in {@link 
#floatToOrderedBytes(float, ByteBuffer)}
+   */
+  public static byte[] doubleToOrderedBytes(double val, ByteBuffer reuse) {
+    ByteBuffer bytes = ByteBuffers.reuse(reuse, Double.BYTES);
+    long lng = Double.doubleToLongBits(val);
+    lng ^= ((lng >> (Long.SIZE - 1)) | Long.MIN_VALUE);
+    bytes.putLong(lng);
+    return bytes.array();
+  }
+
+  /**
+   * Strings are lexicographically sortable BUT if different byte array 
lengths will
+   * ruin the Z-Ordering. (ZOrder requires that a given column contribute the 
same number of bytes every time).
+   * This implementation just uses a set size to for all output byte 
representations. Truncating longer strings
+   * and right padding 0 for shorter strings.
+   */
+  public static byte[] stringToOrderedBytes(String val, int length, ByteBuffer 
reuse) {

Review comment:
       The debate we were having before was wether to limit it to the max 
common length of columns, or whether to let it go beyond that.
   
   Like with (A, B, CC, DDD)
   
   Do you return
   
   A. : All bytes that are available
   ABCDCDD 
   
   B. : All bytes that can be interleaved with at least one other column
   ABCDCD
   
   Or
   
   C. : All bytes that can be interleaved with all other columns
   ABCD
   
   
   My current implementation in Spark just does A which is the sum of all 
column lengths, but we could do B and save some space at the cost of losing a 
bit of single column ordering. I don't think C actually makes a lot of sense 
unless we do some hind of bin hashing and actually generate bytes all of the 
same size.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to