Fwd: Phoenix table stats query performance issue

[Li Gao]

Hi Phoenix community,

We are encountering general performance degradation when doing table stats
query over the leading primary key column in a Phoenix table.

The table schema, used query, JMeter results observed over 1-hour window,
and the python script that slowly generate data into the table.

Please advise how we can improve the performance of the query

Thanks,
Li

import jaydebeapi
conn = jaydebeapi.connect('org.apache.phoenix.jdbc.PhoenixDriver', \
                           ['jdbc:phoenix:phoexix-node-101', '', ''], \
                           './phoenix-4.5.2-HBase-1.0-client.jar')
conn.jconn.setAutoCommit(True)
curs = conn.cursor()
for i in range(0,100000010):
  sql = 'upsert into pk.tests (id, pubid, name) values (%d, %d, \'name %d\')' % (i,i,i) 
  print sql
  curs.execute(sql )
conn.close()
print 'Now all done!'

pk-perf-tables.sql
Description: Binary data