Merge pull request #646 from apache/compressed_iterator

code cleanup
apache · Feb 3, 2025 · 2dd1d5e · 2dd1d5e
2 parents 5753fbd + a0b86ee
commit 2dd1d5e
Show file tree

Hide file tree

Showing 5 changed files with 9 additions and 15 deletions.
diff --git a/src/main/java/org/apache/datasketches/theta/CompactSketch.java b/src/main/java/org/apache/datasketches/theta/CompactSketch.java
@@ -188,8 +188,6 @@ private static CompactSketch wrap(final Memory srcMem, final long seed, final bo
     final short seedHash = ThetaUtil.computeSeedHash(seed);
 
     if (serVer == 4) {
-      // not wrapping the compressed format since currently we cannot take advantage of
-      // decompression during iteration because set operations reach into memory directly
       return DirectCompactCompressedSketch.wrapInstance(srcMem,
           enforceSeed ? seedHash : (short) extractSeedHash(srcMem));
     }
@@ -251,6 +249,11 @@ public boolean isCompact() {
     return true;
   }
 
+  @Override
+  public double getEstimate() {
+    return Sketch.estimate(getThetaLong(), getRetainedEntries());
+  }
+
   /**
    * gets the sketch as a compressed byte array
    * @return the sketch as a compressed byte array

diff --git a/src/main/java/org/apache/datasketches/theta/DirectCompactCompressedSketch.java b/src/main/java/org/apache/datasketches/theta/DirectCompactCompressedSketch.java
@@ -118,7 +118,8 @@ public boolean isOrdered() {
   public HashIterator iterator() {
     return new MemoryCompactCompressedHashIterator(
       mem_,
-      (extractPreLongs(mem_) > 1 ? 16 : 8) + extractNumEntriesBytesV4(mem_),
+      (extractPreLongs(mem_) > 1 ? START_PACKED_DATA_ESTIMATION_MODE : START_PACKED_DATA_EXACT_MODE)
+        + extractNumEntriesBytesV4(mem_),
       extractEntryBitsV4(mem_),
       getRetainedEntries()
     );

diff --git a/src/main/java/org/apache/datasketches/theta/DirectCompactSketch.java b/src/main/java/org/apache/datasketches/theta/DirectCompactSketch.java
@@ -84,11 +84,6 @@ public int getCurrentBytes() {
     return (preLongs + curCount) << 3;
   }
 
-  @Override
-  public double getEstimate() {
-    return Sketch.estimate(getThetaLong(), getRetainedEntries());
-  }
-
   @Override
   public int getRetainedEntries(final boolean valid) { //compact is always valid
     if (otherCheckForSingleItem(mem_)) { return 1; }

diff --git a/src/main/java/org/apache/datasketches/theta/HeapCompactSketch.java b/src/main/java/org/apache/datasketches/theta/HeapCompactSketch.java
@@ -87,11 +87,6 @@ public int getCurrentBytes() {
     return (preLongs_ + curCount_) << 3;
   }
 
-  @Override
-  public double getEstimate() {
-    return Sketch.estimate(thetaLong_, curCount_);
-  }
-
   @Override
   public int getRetainedEntries(final boolean valid) {
     return curCount_;

diff --git a/src/main/java/org/apache/datasketches/theta/MemoryCompactCompressedHashIterator.java b/src/main/java/org/apache/datasketches/theta/MemoryCompactCompressedHashIterator.java
@@ -23,8 +23,8 @@
 
 import org.apache.datasketches.memory.Memory;
 
-/**
- * @author Lee Rhodes
+/*
+ * This is to uncompress serial version 4 sketch incrementally
  */
 class MemoryCompactCompressedHashIterator implements HashIterator {
   private Memory mem;