Phoenix JDBC driver too slow

Question

I am using phoenix driver 4.0 version to query database. I have also tried setting hbase.client.scanner.caching to 1000 and 10000 In all the cases it takes 3 minutes to iterate over 10000 rows. I have not put any code inside iteration loop. I have to scan over 1L records so it is taking more than half hour.

What can be the reason for this.

long count = 0;
            while (rs.next())
            {
                if (count % 1000 == 0)
                {
                    log.info("Count == " + count);
                }
                count++;
            }

What is your row key design? your records spread across multiple regions with salting and hashing done? — Ram Ghadiyaram, Apr 28 '16 at 07:07

score 0 · Answer 1 · answered Dec 27 '16 at 14:00

I think hadoop and its component are good at group-like function,and hash distribute record by using rowkey to find record quickly.but not good at iterate,you run 10000 record maybe cost the same as 100000 and even 1000000.I hope you can get what mean I say through my poor English.:)

Phoenix JDBC driver too slow

1 Answers1