We are using Cognos to read Microsoft Analysis Services cubes. We have a dimension with a 2 byte code and a 20 byte descriptions. There is a one-to-one correspondence between a code and a description, and we don't want to have two separate levels in the hierarchy--one for the code and one for the description. Cognos can only read what is passed to it, and can't use substring or any other character functions. When passing the full 20 byte description to a query, performance suffers. It improves dramatically when we use the 2 byte code. Is there a way to display the description on the Cognos report, but use the code when querying against the cube to improve performance?
Yup. In your Transformer map, double-click on the field that's currently building the dimension, and there'll be a window with some tabs, should be on the 1st tab, 'Source'. Then, look under 'associations'. You'll have only the 'source', with your dimension 20-byte description. Open this folder for 'Source', and pick the 2-byte code. Click the 'unique' checkbox, and answer 'yes'. Then open the next folder 'Label' and pick the 20-byte description. The effect of this is to build the cube using the 2-byte code and have the 20-byte description in the cube folder. The best analogy for this would be when u build prompts in Impromptu, putting the code & description fields in 'Use' & 'Display', respectively. Performance should improve
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.