[SPARK-36326] Use Map.computeIfAbsent to simplify the process of HeapMemoryAllocator.bufferPoolsBySize init new item

• Type: Improvement
• Status: Resolved
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 3.3.0
• Fix Version/s: 3.3.0
• Component/s: Spark Core
• Assignee: Yang Jie
• Reporter: Yang Jie
• Created: 28/Jul/21 08:45
• Updated: 29/Jul/21 15:18
• Resolved: 29/Jul/21 15:18

Description

HeapMemoryAllocator initializes the new element ofbufferPoolsBySizein the following way：

synchronized (this) {
  LinkedList<WeakReference<long[]>> pool = bufferPoolsBySize.get(alignedSize);
  if (pool == null) {
    pool = new LinkedList<>();
    bufferPoolsBySize.put(alignedSize, pool);
  }
  pool.add(new WeakReference<>(array));
}

The codes can be simplified by Map.computeIfAbsent

 

 

Issue Links

• links to: [Github] Pull Request #33558 (LuciferYang)