Database Reference
In-Depth Information
rolls up market updating six blocks. This section performed 33 reads and six writes.
using the log output, there are two ways to determine if all of the appropriate blocks
have been aggregated. First, the outline can be compared to the blocks outputted to
the log file. A second and easier method is to change the sales number for “100-10” and
oregon, followed by “Calc All” with intelligent calculation turned on. The log will show
every block that Essbase thinks it needs to change. note that the read and write counts
may differ slightly because Essbase processes the blocks a little more efficiently.
4 .7. 6
Create Blocks Explained
missing data blocks in BSo can be perplexing. For anyone new to Essbase, the concept
of missing blocks is hard to grasp. Even experienced developers and administrators find
block creation confusing. I expect that most of the readers have experienced situations
when blocks must be explicitly created. For those who have not, this section might not
make a lot of sense, but eventually you will need these tips.
The most important thing to remember when creating blocks is that the goal is to
create the appropriate block without creating extra blocks. As mentioned repeatedly
in this chapter, the more blocks that are processed, the slower the database. It is for
this reason that the more experienced developer's tend to shy away from the statements
CrEAtEBLoCkonEQ and CrEAtEnonmISSIngBLk. While these statements are
powerful, they can needlessly increase the number of blocks if not used correctly.
he commands CrEAtEBLoCkonEQ and CrEAtEnonmISSIngBLk have not
always been available in Essbase. Prior to their introduction, blocks were created by
rearranging the formula so that the result is placed into a sparse member. The following
script will not create a block because Sales is a member of a dense dimension.
This script assumes that the block for Cola->nevada->Actual did not exist prior to
running the script, as shown in table 4.4.
FIX (Cola, Nevada, Actual)
SET CREATEBLOCKONEQ OFF;
Sales = "NEW YORK" + 100;
ENDFIX
rewriting the formula so that the result is placed in nevada, which is sparse, creates
the necessary block, as shown in table 4.5.
Table 4.4
Data Results
Sales
Jan
Actual
Budget
Cola
nevada
#mi
#mi
n e w yo r k
25
#mi
Table 4.5
Data Results after Change
Sales
Jan
Actual
Budget
Cola
nevada
125
#mi
n e w yo r k
25
#mi
