Binary allocator

cpp/pom.xml

@@ -22,7 +22,7 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-parent</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>tsfile-cpp</artifactId>
     <packaging>pom</packaging>

java/common/pom.xml

@@ -24,7 +24,7 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-java</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>common</artifactId>
     <name>TsFile: Java: Common</name>

java/common/src/main/java/org/apache/tsfile/utils/Binary.java

@@ -36,7 +36,7 @@ public class Binary implements Comparable<Binary>, Serializable, Accountable {
   private static final long serialVersionUID = 6394197743397020735L;
   public static final Binary EMPTY_VALUE = new Binary(new byte[0]);
 
-  private byte[] values;
+  byte[] values;
 
   /** if the bytes v is modified, the modification is visible to this binary. */
   public Binary(byte[] v) {
@@ -70,7 +70,7 @@ public int compareTo(Binary other) {
   }
 
   // avoid overflow
-  private char getChar(byte[] val, int index) {
+  protected char getChar(byte[] val, int index) {
     return (char) (val[index] & 0xff);
   }
 
@@ -92,9 +92,13 @@ public int hashCode() {
   }
 
   /**
-   * get length.
+   * Gets the actual payload length.
    *
-   * @return length
+   * <p>This method returns the effective length of the data (payload) stored in the byte array. If
+   * the byte array is null, it returns -1. Note that this length may be less than the total
+   * capacity of the byte array.
+   *
+   * @return the actual payload length, or -1 if the byte array is null
    */
   public int getLength() {
     if (this.values == null) {
@@ -103,6 +107,22 @@ public int getLength() {
     return this.values.length;
   }
 
+  /**
+   * Gets the total capacity of the byte array.
+   *
+   * <p>This method returns the total capacity of the underlying byte array. If the byte array is
+   * null, it returns -1. Note that the effective payload length (actual valid data) may be smaller
+   * than the total capacity of the array.
+   *
+   * @return the total capacity of the byte array, or -1 if the array is null
+   */
+  public int getCapacity() {
+    if (this.values == null) {
+      return -1;
+    }
+    return this.values.length;
+  }
+
   public String getStringValue(Charset charset) {
     return new String(this.values, charset);
   }
@@ -117,6 +137,10 @@ public byte[] getValues() {
     return values;
   }
 
+  public Pair<byte[], Integer> getValuesAndLength() {
+    return new Pair<>(values, values.length);
+  }
+
   public void setValues(byte[] values) {
     this.values = values;
   }
@@ -125,4 +149,12 @@ public void setValues(byte[] values) {
   public long ramBytesUsed() {
     return INSTANCE_SIZE + sizeOf(values);
   }
+
+  public long ramShallowBytesUsed() {
+    return INSTANCE_SIZE;
+  }
+
+  public boolean isNull() {
+    return values == null;
+  }
 }

java/common/src/main/java/org/apache/tsfile/utils/PooledBinary.java

@@ -18,11 +18,11 @@
  */
 package org.apache.tsfile.utils;
 
-import java.io.Serializable;
 import java.nio.charset.Charset;
 import java.nio.charset.StandardCharsets;
 
 import static org.apache.tsfile.utils.RamUsageEstimator.shallowSizeOfInstance;
+import static org.apache.tsfile.utils.RamUsageEstimator.sizeOf;
 
 /**
  * This class represents a pooled binary object for application layer. It is designed to improve
@@ -31,64 +31,32 @@
  * length of the underlying byte array. Always use getLength() instead of getValue().length to get
  * the correct length.
  */
-public class PooledBinary implements Comparable<PooledBinary>, Serializable, Accountable {
+public class PooledBinary extends Binary {
 
   private static final long INSTANCE_SIZE = shallowSizeOfInstance(PooledBinary.class);
   private static final long serialVersionUID = 6394197743397020735L;
-  public static final PooledBinary EMPTY_VALUE = new PooledBinary(new byte[0]);
-
-  private Binary binary;
 
   private int length;
 
   private int arenaIndex = -1;
 
   /** if the bytes v is modified, the modification is visible to this binary. */
   public PooledBinary(byte[] v) {
-    this.binary = new Binary(v);
-    this.length = binary.getLength();
+    super(v);
+    this.length = values.length;
   }
 
   public PooledBinary(String s, Charset charset) {
-    this.binary = new Binary(s, charset);
-    this.length = binary.getLength();
+    super(s, charset);
+    this.length = values.length;
   }
 
   public PooledBinary(byte[] v, int length, int arenaIndex) {
-    this.binary = new Binary(v);
+    super(v);
     this.length = length;
     this.arenaIndex = arenaIndex;
   }
 
-  @Override
-  public int compareTo(PooledBinary other) {
-    if (other == null) {
-      if (this.binary.getValues() == null) {
-        return 0;
-      } else {
-        return 1;
-      }
-    }
-
-    // copied from StringLatin1.compareT0
-    int len1 = getLength();
-    int len2 = other.getLength();
-    int lim = Math.min(len1, len2);
-    byte[] v0 = this.binary.getValues();
-    byte[] v1 = other.binary.getValues();
-    for (int k = 0; k < lim; k++) {
-      if (v0[k] != v1[k]) {
-        return getChar(v0, k) - getChar(v1, k);
-      }
-    }
-    return len1 - len2;
-  }
-
-  // avoid overflow
-  private char getChar(byte[] val, int index) {
-    return (char) (val[index] & 0xff);
-  }
-
   @Override
   public boolean equals(Object o) {
     if (this == o) {
@@ -103,8 +71,8 @@ public boolean equals(Object o) {
       return false;
     }
 
-    byte[] v0 = binary.getValues();
-    byte[] v1 = otherBinary.binary.getValues();
+    byte[] v0 = values;
+    byte[] v1 = otherBinary.values;
 
     for (int i = 0; i < length; i++) {
       if (v0[i] != v1[i]) {
@@ -118,26 +86,24 @@ public boolean equals(Object o) {
   @Override
   public int hashCode() {
     // copied from Arrays.hashCode
-    if (binary.getValues() == null) return 0;
+    if (values == null) return 0;
 
     int result = 1;
-    byte[] val = binary.getValues();
-    for (int i = 0; i < length; i++) result = 31 * result + val[i];
+    for (int i = 0; i < length; i++) {
+      result = 31 * result + values[i];
+    }
 
     return result;
   }
 
-  /**
-   * get length.
-   *
-   * @return length
-   */
+  @Override
   public int getLength() {
     return this.length;
   }
 
+  @Override
   public String getStringValue(Charset charset) {
-    return new String(this.binary.getValues(), 0, length, charset);
+    return new String(values, 0, length, charset);
   }
 
   @Override
@@ -146,17 +112,19 @@ public String toString() {
     return getStringValue(StandardCharsets.UTF_8);
   }
 
-  public byte[] getValues() {
-    return binary.getValues();
+  @Override
+  public Pair<byte[], Integer> getValuesAndLength() {
+    return new Pair<>(values, length);
   }
 
+  @Override
   public void setValues(byte[] values) {
-    this.binary.setValues(values);
-    this.length = this.binary.getLength();
+    super.setValues(values);
+    this.length = values.length;
   }
 
   public void setValues(byte[] values, int length) {
-    this.binary.setValues(values);
+    super.setValues(values);
     this.length = length;
   }
 
@@ -166,10 +134,11 @@ public int getArenaIndex() {
 
   @Override
   public long ramBytesUsed() {
-    return INSTANCE_SIZE + binary.ramBytesUsed();
+    return INSTANCE_SIZE + sizeOf(values);
   }
 
-  public Binary toBinary() {
-    return binary;
+  @Override
+  public long ramShallowBytesUsed() {
+    return INSTANCE_SIZE;
   }
 }

java/examples/pom.xml

@@ -24,7 +24,7 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-java</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>examples</artifactId>
     <packaging>pom</packaging>
@@ -37,7 +37,7 @@
         <dependency>
             <groupId>org.apache.tsfile</groupId>
             <artifactId>tsfile</artifactId>
-            <version>1.2.0-SNAPSHOT</version>
+            <version>1.3.0-SNAPSHOT</version>
         </dependency>
     </dependencies>
     <build>

java/pom.xml

@@ -24,10 +24,10 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-parent</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>tsfile-java</artifactId>
-    <version>1.2.0-SNAPSHOT</version>
+    <version>1.3.0-SNAPSHOT</version>
     <packaging>pom</packaging>
     <name>TsFile: Java</name>
     <modules>

java/tools/pom.xml

@@ -24,15 +24,15 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-java</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>tools</artifactId>
     <name>TsFile: Java: Tools</name>
     <dependencies>
         <dependency>
             <groupId>org.apache.tsfile</groupId>
             <artifactId>common</artifactId>
-            <version>1.2.0-SNAPSHOT</version>
+            <version>1.3.0-SNAPSHOT</version>
         </dependency>
         <dependency>
             <groupId>commons-cli</groupId>
@@ -50,7 +50,7 @@
         <dependency>
             <groupId>org.apache.tsfile</groupId>
             <artifactId>tsfile</artifactId>
-            <version>1.2.0-SNAPSHOT</version>
+            <version>1.3.0-SNAPSHOT</version>
         </dependency>
         <dependency>
             <groupId>ch.qos.logback</groupId>

java/tsfile/pom.xml

@@ -24,7 +24,7 @@
     <parent>
         <groupId>org.apache.tsfile</groupId>
         <artifactId>tsfile-java</artifactId>
-        <version>1.2.0-SNAPSHOT</version>
+        <version>1.3.0-SNAPSHOT</version>
     </parent>
     <artifactId>tsfile</artifactId>
     <name>TsFile: Java: TsFile</name>
@@ -38,7 +38,7 @@
         <dependency>
             <groupId>org.apache.tsfile</groupId>
             <artifactId>common</artifactId>
-            <version>1.2.0-SNAPSHOT</version>
+            <version>1.3.0-SNAPSHOT</version>
         </dependency>
         <dependency>
             <groupId>com.github.luben</groupId>

java/tsfile/src/main/java/org/apache/tsfile/encoding/encoder/DictionaryEncoder.java

@@ -21,6 +21,7 @@
 
 import org.apache.tsfile.file.metadata.enums.TSEncoding;
 import org.apache.tsfile.utils.Binary;
+import org.apache.tsfile.utils.BinaryUtils;
 import org.apache.tsfile.utils.ReadWriteForEncodingUtils;
 
 import org.slf4j.Logger;
@@ -99,7 +100,7 @@ private void writeMap(ByteArrayOutputStream out) throws IOException {
     ReadWriteForEncodingUtils.writeVarInt(indexEntry.size(), out);
     for (Binary value : indexEntry) {
       ReadWriteForEncodingUtils.writeVarInt(value.getLength(), out);
-      out.write(value.getValues());
+      BinaryUtils.serializeBytes(out, value);
     }
   }
 

java/tsfile/src/main/java/org/apache/tsfile/encoding/encoder/PlainEncoder.java

@@ -24,6 +24,7 @@
 import org.apache.tsfile.exception.encoding.TsFileEncodingException;
 import org.apache.tsfile.file.metadata.enums.TSEncoding;
 import org.apache.tsfile.utils.Binary;
+import org.apache.tsfile.utils.BinaryUtils;
 import org.apache.tsfile.utils.ReadWriteForEncodingUtils;
 
 import org.slf4j.Logger;
@@ -92,7 +93,7 @@ public void encode(Binary value, ByteArrayOutputStream out) {
       // write the length of the bytes
       encode(value.getLength(), out);
       // write value
-      out.write(value.getValues());
+      BinaryUtils.serializeBytes(out, value);
     } catch (IOException e) {
       logger.error(
           "tsfile-encoding PlainEncoder: error occurs when encode Binary value {}", value, e);

java/tsfile/src/main/java/org/apache/tsfile/file/metadata/statistics/BinaryStatistics.java

@@ -54,7 +54,7 @@ public TSDataType getType() {
   /** The output of this method should be identical to the method "serializeStats(outputStream)". */
   @Override
   public int getStatsSize() {
-    return 4 + firstValue.getValues().length + 4 + lastValue.getValues().length;
+    return 4 + firstValue.getLength() + 4 + lastValue.getLength();
   }
 
   @Override