
Basic Concepts Brio 2002
Copyright © 2002, Brio Software Page 5
Sometimes, however, you want to compare one consolidation with another. In
these cases, you have several ways to accomplish this. You can use Computed
Items that are created at the Chart or Pivot section itself if the comparison is to be
used only in that chart or pivot. To create globally useful consolidation
comparisons, start in the Results or Table section with Computed Items that
summarize at the relevant group levels, then add another Computed Item to
calculate the actual comparison. When you add this comparison item to a chart
or pivot, change the data function from Sum to something appropriate to
eliminate multiple counting.
Local versus Server Processing
People often overlook the power of their database to serve their reporting and
analysis purposes. For example, one retailer wanted to graph the number and
dollar amount of transactions per department on an hour-by-hour basis. They
wanted to use this chart to estimate required staffing levels throughout the day
based on actual sales history.
The original query processing and chart generation took over 2 hours. This kind
of response was clearly unacceptable. Two factors contributed to the amount of
processing time. First, of course, they were fetching every single record of sales
data for the given period, amounting to more than 100,000 rows. Then, Brio had
to process all those rows to produce the graphical consolidation that was the
desired end result.
As a general rule, anything you do on the Query section is processed at the
database. Actions performed on any other section are executed by Brio itself. I
showed them how they could speed up the processing by having their database
perform the consolidation of the data instead of having Brio do it. They only had
to make a minor change to the query.
By default, Brio creates a simple Select statement from the Request line in the
Query section. You can change this behavior by selecting the value(s) on the
Request Line you want to consolidate, right-clicking and selecting the Data
Function item from the speed menu which appears.
In this case, I had the retailer select the Sum function. This time, the entire
process returned 52 rows and was finished in less than one minute.
Data Presentation Requirements
You also need to look at how you will present the data to your users. Drilling
options will determine how you build the data model. Human factors
engineering concepts affect which pivot or chart display options you use.
Drilling Requirements
If you want to use certain drilling options, they must be predefined in the data
model itself. The ability to use pre-defined drill paths, drill-to-detail and
incremental drilling all depend on the data model used by the query. They are
not controlled in the individual chart or pivot. The ability to turn off drilling
altogether is also found in the data model definition.
Comentarios a estos manuales