GH-3522: Reuse intermediate buffers in RunLengthBitPackingHybridDecoder PACKED path

Allocate the int[] values buffer and byte[] read-staging buffer once per
decoder and grow them lazily, instead of allocating fresh arrays on every
PACKED run. Resolves the existing "TODO: reuse a buffer" comment.

A new currentBufferLength field tracks the logical length of the active
region in packedValuesBuffer (which may now exceed the current run's
size after a prior larger run grew it).

Benchmark (RleDictionaryIndexDecodingBenchmark, 100k INT32, BIT_WIDTH=10,
JMH -wi 5 -i 10 -f 2):

  Pattern         | master ops/s | optimized ops/s | Improvement
  SEQUENTIAL      |  93,061,521  |   113,856,860   |   +22.3%
  RANDOM          |  92,929,824  |   114,238,638   |   +22.9%
  LOW_CARDINALITY |  92,813,229  |   115,271,347   |   +24.2%

End-to-end FileReadBenchmark sees ~2% improvement (RLE decoding is a
small fraction of full file reads).

Validation: 573 parquet-column tests pass. Built with
-Dspotless.check.skip=true -Drat.skip=true -Djapicmp.skip=true.

Author: iemejia

parquet-column/src/main/java/org/apache/parquet/column/values/rle/RunLengthBitPackingHybridDecoder.java

@@ -48,6 +48,11 @@ private static enum MODE {
   private int currentCount;
   private int currentValue;
   private int[] currentBuffer;
+  private int currentBufferLength;
+
+  // Reusable buffers to avoid per-run allocation in PACKED mode
+  private int[] packedValuesBuffer = new int[0];
+  private byte[] packedBytesBuffer = new byte[0];
 
   public RunLengthBitPackingHybridDecoder(int bitWidth, InputStream in) {
     LOG.debug("decoding bitWidth {}", bitWidth);
@@ -69,7 +74,7 @@ public int readInt() throws IOException {
         result = currentValue;
         break;
       case PACKED:
-        result = currentBuffer[currentBuffer.length - 1 - currentCount];
+        result = currentBuffer[currentBufferLength - 1 - currentCount];
         break;
       default:
         throw new ParquetDecodingException("not a valid mode " + mode);
@@ -90,17 +95,24 @@ private void readNext() throws IOException {
       case PACKED:
         int numGroups = header >>> 1;
         currentCount = numGroups * 8;
+        currentBufferLength = currentCount;
         LOG.debug("reading {} values BIT PACKED", currentCount);
-        currentBuffer = new int[currentCount]; // TODO: reuse a buffer
-        byte[] bytes = new byte[numGroups * bitWidth];
+        if (packedValuesBuffer.length < currentCount) {
+          packedValuesBuffer = new int[currentCount];
+        }
+        currentBuffer = packedValuesBuffer;
+        int bytesRequired = numGroups * bitWidth;
+        if (packedBytesBuffer.length < bytesRequired) {
+          packedBytesBuffer = new byte[bytesRequired];
+        }
         // At the end of the file RLE data though, there might not be that many bytes left.
         int bytesToRead = (int) Math.ceil(currentCount * bitWidth / 8.0);
         bytesToRead = Math.min(bytesToRead, in.available());
-        new DataInputStream(in).readFully(bytes, 0, bytesToRead);
+        new DataInputStream(in).readFully(packedBytesBuffer, 0, bytesToRead);
         for (int valueIndex = 0, byteIndex = 0;
             valueIndex < currentCount;
             valueIndex += 8, byteIndex += bitWidth) {
-          packer.unpack8Values(bytes, byteIndex, currentBuffer, valueIndex);
+          packer.unpack8Values(packedBytesBuffer, byteIndex, currentBuffer, valueIndex);
         }
         break;
       default: