Difference between details/dimensions

[Guest_imported]

Hi All,
I am analyzing data on a pretty huge d/b, 1.7 million records.. When I create a new var as a dimension/detail, there seem to be a speed difference in the execution time (its taking ages anyway ~ 30 minutes) to process.. I was hoping someone could enlighten me when/why to use dimension/detail appropriately.. thanks

 

[blom0344]

Hello Asok8,

If you would compare the SQL between objects defined as dimensions or details you would see no difference. The actual statement is the same. In fact query speed should be independent of whether you are retrieving dimensions or details. Defining objects as dimensions has a couple of reasons:

1. Use them in multidimensional analysis ; drilling
2. Use them to link datasets at report level

You can regard details as the properties of dimensions. So, if the customernumber is the dimension, the customername is a detail. Details are inherently non-unique, dimension-objects mostly unique.

Creating new document variables dimension or detail will prompt BO do to a reformatting of the document. There are 3 ways to deal with that:

1. Work on a document in the structure view, no data is shown, so formatting is much faster.
2. Work on a limited subset when you build a report by temporarily setting a condition that will limit the amount of records retrieved.
3. Apply a global filter to diminish the amount of data shown.

Last remark:

Some reports can be dramatically speeded up by creating aggregating measures in the universe (like sum(sales), so instead of retrieving a zillion records, you get of fraction of the number of rows. The reformatting is then quite a bit quicker. T. Blom
Information analist
Shimano Europe
tbl@shimano-eu.com