Samarth Jain created PHOENIX-1648:
-------------------------------------
Summary: Extra scan being issued while doing SELECT COUNT(*)
queries
Key: PHOENIX-1648
URL: https://issues.apache.org/jira/browse/PHOENIX-1648
Project: Phoenix
Issue Type: Bug
Affects Versions: 5.0.0, 4.3
Reporter: Samarth Jain
On turning tracing on and executing SELECT COUNT(*) queries, I am seeing an
extra scan being executed every time.
CREATE TABLE MY_TABLE (ID INTEGER NOT NULL PRIMARY KEY, VALUE INTEGER)
SALT_BUCKETS = 16
SELECT COUNT(*) FROM MY_TABLE
The trace table has:
Creating basic query for [CLIENT 16-CHUNK PARALLEL 16-WAY FULL SCAN OVER
MY_TABLE, SERVER FILTER BY FIRST KEY ONLY, SERVER AGGREGATE INTO SINGLE
ROW]
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Parallel scanner for table: MY_TABLE
Creating basic query for [CLIENT 1-CHUNK PARALLEL 1-WAY RANGE SCAN OVER
SYSTEM.CATALOG [null,null,'MY_TABLE',not null], SERVER FILTER BY
COLUMN_FAMILY IS NULL]
Parallel scanner for table: SYSTEM.CATALOG
While the 16 scanners being created for MY_TABLE is expected, the extra scanner
for SYSTEM.CATALOG isn't. This is happening consistently, so this likely isn't
happening because of cache expiration.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
