[HBASE-25576] Should not use a byte array as a key of HashMap or an element of HashSet

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Labels: beginner
• Assignee: Unassigned
• Reporter: Toshihiro Suzuki
• Created: 13/Feb/21 13:16
• Updated: 01/Mar/23 14:59

Description

I sometimes face the code where we use a byte array as a key of HashMap (ConcurrentHashMap) or an element of HashSet, which could cause very confusing bugs.

The following code is an example where we use a byte array as a key of HashMap:
https://github.com/apache/hbase/blob/326835e8372cc83092e0ec127650438ff153476a/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/wal/SequenceIdAccounting.java#L94

In Java, we can't use a byte array as a key of HashMap or an element of HashSet directly. One solution for it is to use TreeMap (ConcurrentSkipListMap) or TreeSet (ConcurrentSkipListSet) with Bytes.BYTES_COMPARATOR instead of HashMap and HashSet as follows:

Map<byte[], String> map1 = new TreeMap<>(Bytes.BYTES_COMPARATOR);
Map<byte[], String> map2 = new ConcurrentSkipListMap<>(Bytes.BYTES_COMPARATOR);

Set<byte[]> set1 = new TreeSet<>(Bytes.BYTES_COMPARATOR);
Set<byte[]> set2 = new ConcurrentSkipListSet<>(Bytes.BYTES_COMPARATOR);

We should fix the existing ones in this Jira.