Skip to Content

Does my original query execute even when not used in design studio charts?

Dear Gurus,

We have got a huge query with almost 250 measures and 30 dimensions.

Our understanding is that if we bring this query in design studio and create multiple small initial views of the same query and use them to map with our layout then it would improve the performance.

But the loading time of the dashboard is still the same as using the original (huge) structure of the query and there is no improvement in our performance.

Should we split the original query and have small subset of the queries or should we still continue with the initial views?

Also, can we use initial views of the query for parallel processing if it is the same query?



Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

2 Answers

  • Jul 25, 2017 at 12:03 AM

    When Design Studio request the data it will be based on what is set in the initial view of the datasource, if the datasource is defined to have only 2 dimensions and 10 measures it will only request for those. But if the 10 measures requested requires other measures for doing internal calculations or if there is Exception Aggregation defined which require reading the data at the most detail level it will have to get all the required measures and dimensions to get the required output.

    So splitting the query alone wont help in your case. Try to run the query with the datasource required for design studio from RSRT2 and see whether the performance is comparable, make sure the dimension used in the datasource is same as the dimensions in the RSRT2 output. Also when you have multiple datasources from a Design Studio standpoint whether it is the same query or not does not matter, each of them are requested seperately from the backend unless you have Parallel processing ON.

    See note below for more trouble shooting.

    Add comment
    10|10000 characters needed characters exceeded

  • Jul 24, 2017 at 01:41 PM

    Hi Sumit,

    To start with what version BO and DS business is running on?

    What is your observation with performance statistics with current intake of 250M to 30D with in the query ? (Still in my my opinion its massive and will take performance hit)

    Idea of one using single query is great but one major factor is the data. Do you know the row or amount of data this query will be pulling-in?



    Add comment
    10|10000 characters needed characters exceeded

    • Hi Sumit,

      Deepu has provided some good insights in his answer above. Regarding your question about the effectiveness of parallel processing for your particular case, I think it is best to simply try it and then proceed accordingly based on the results.