Quantcast
Channel: SCN : Unanswered Discussions - Semantic Layer
Viewing all articles
Browse latest Browse all 1291

Removing an aggregate for a dimension

$
0
0

Hello all,

 

I am new to BI and I was asked to help fix a performance issue for one of the reports.  One of the things that i noticed is that the SQL generated has a max on a nvarchar string (200 characters long) so the performance for the overall report is very bad (> 9 min for 500K recs), but if I execute the script by hand as seen in SCRIPT2, it takes less than a minute.

 

--SCRIPT2

SELECT

  sum(Table2.Item_count),

  Table1.NAME,

  Table1.ID

FROM

  Table2

           INNER JOIN Table1 ON (Table1.ID = Table2.IntID)

GROUP BY

  Table1.ID,

  Table1.NAME

 

 

Below is a simplified version of the SQL script generated from the .UNV & .UNX:

 

--SCRIPT1

SELECT

  sum(Table2.Item_count),

  MAX( Table1.NAME ),

  Table1.ID

FROM

  Table2

           INNER JOIN Table1 ON (Table1.ID = Table2.IntID)

GROUP BY

  Table1.ID

 

My question is can I change a universe parameter to change the default aggregation of the string to not aggregate, but to place it into the group by as seen in SCRIPT2 as this impacts many other reports.

 

Thanks in advance for any help and suggestions.


Viewing all articles
Browse latest Browse all 1291

Trending Articles