Quantcast
Channel: SCN : Discussion List - SAP HANA and In-Memory Computing
Viewing all articles
Browse latest Browse all 5653

JOIN performance

$
0
0

Hello.

 

I deploy TPC-H test database (scale factor 50) on my test virtual machine (SLES 11SP3, 16VCPU, 48GB RAM)

on HANA SPS09 rev. 96 database.

 

Linear fullscan performance is amazing, but for query's with join's operation got some not optimal performance

 

For example (little changed  3.sql  script from TPC-H) :

 

select  TOP 100        l_orderkey,        sum(l_extendedprice * (1 - l_discount)) as revenue,        o_orderdate,        o_shippriority
from       tpchusr.customer,       tpchusr.orders,       tpchusr.lineitem
where        c_custkey = o_custkey        and l_orderkey = o_orderkey         and c_mktsegment = 'AUTOMOBILE'
group by        l_orderkey,        o_orderdate,        o_shippriority
order by        revenue desc,        o_orderdate   

Tables rows count :

 

LINEITEM - 300 M

ORDERS - 75 M

CUSTOMER 7.5 M

 

Sustained average execution time of this query is 35 sec.

 

I try to create indexes on join columns, create statistics on all tables, use some JOIN-hint's, but with no luck

 

I collect and attach EXPLAIN_PLAN, PlanViz graph, and output of  "HANA_SQL_StatementHash_KeyFigures.txt"  script.

 

Will be very appricated if someone can give any ideas for improvment of this query

 

Thank's


Viewing all articles
Browse latest Browse all 5653

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>