Gokcen Iskender created PHOENIX-5652:
----------------------------------------
Summary: IndexScrutinyTool perf improvement to query indexes using
skipscanfilter
Key: PHOENIX-5652
URL: https://issues.apache.org/jira/browse/PHOENIX-5652
Project: Phoenix
Issue Type: Improvement
Reporter: Gokcen Iskender
When the IndexScrutinyTool uses DATA_TABLE as source (source type DATA_TABLE or
BOTH), it constructs a query that uses a WHERE IN statement to look up the data
table entries in the index table.
Consider a table with ID Integer PK, NAME varchar and index with PK NAME
The current query is constructed like this:
IndexTable WHERE (:ID, 0:NAME) IN ()
The WhereOptimizer looks at the column order to decide whether to use
SkipScanFilter or not. The PK of the index table should come before the PK of
data table.
If the query is written like this:
IndexTable WHERE (0:NAME, :ID) IN () then SkipScanFilter is used and the query
is much faster.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
