[orc] branch main updated: ORC-833: Calculate nextVector Batch Size Once (#738)

[do...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

dongjoon pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/orc.git


The following commit(s) were added to refs/heads/main by this push:
     new 7566d57  ORC-833: Calculate nextVector Batch Size Once (#738)
7566d57 is described below

commit 7566d57e8a71eed8f802026c84be734e9152f565
Author: belugabehr <12...@users.noreply.github.com>
AuthorDate: Wed Jul 7 17:18:38 2021 -0400

    ORC-833: Calculate nextVector Batch Size Once (#738)
    
    ### What changes were proposed in this pull request?
    
    The number of iterations of the loops in RunLengthIntegerReaderV2 nextVector look at two distinct, but known, values. Instead of check each value in each loop, pre-compute the batch size.
    
    ### Why are the changes needed?
    
    Code clarity, and if the optimizer doesn't already do it automagically, saves a few branches.
    
    ### How was this patch tested?
    
    No changes in functionality.  Uses existing unit tests.
---
 .../org/apache/orc/impl/RunLengthIntegerReaderV2.java     | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)

diff --git a/java/core/src/java/org/apache/orc/impl/RunLengthIntegerReaderV2.java b/java/core/src/java/org/apache/orc/impl/RunLengthIntegerReaderV2.java
index aaed5db..bf96d2b 100644
--- a/java/core/src/java/org/apache/orc/impl/RunLengthIntegerReaderV2.java
+++ b/java/core/src/java/org/apache/orc/impl/RunLengthIntegerReaderV2.java
@@ -390,20 +390,15 @@ public class RunLengthIntegerReaderV2 implements IntegerReader {
   }
 
   @Override
-  public void nextVector(ColumnVector vector,
-                         int[] data,
-                         int size) throws IOException {
+  public void nextVector(ColumnVector vector, int[] data, int size) throws IOException {
+    final int batchSize = Math.min(data.length, size);
     if (vector.noNulls) {
-      for(int r=0; r < data.length && r < size; ++r) {
+      for (int r = 0; r < batchSize; ++r) {
         data[r] = (int) next();
       }
     } else if (!(vector.isRepeating && vector.isNull[0])) {
-      for(int r=0; r < data.length && r < size; ++r) {
-        if (!vector.isNull[r]) {
-          data[r] = (int) next();
-        } else {
-          data[r] = 1;
-        }
+      for (int r = 0; r < batchSize; ++r) {
+        data[r] = (vector.isNull[r]) ? 1 : (int) next();
       }
     }
   }