Oracle Essbase 9 Implementation Guide- P38 pptx

For starters, this is how you would code the simplest of FIX and ENDFIX commands: FIX@IDESCENDENTS"Calendar Periods" "Gross Stock" = "Production" – "Sales"; ENDFIX Notice how the FIX co

Do this instead:

FIX(@IDESCENDANTS("Calendar Periods" AND

@LEVMBRS("Calendar Periods",0))

Any Calculation Function;

ENDFIX

In the following examples, we will show you how to narrow down the amount

of data you will calculate in your database for various scenarios

Basic FIX and ENDFIX examples

To begin with, as mentioned earlier, FIX and ENDFIX commands can be nested We

have also given you some tips regarding when to use the FIX and ENDFIX commands and what to avoid when nesting them, like the same dimensions named in separate FIX commands

For starters, this is how you would code the simplest of FIX and ENDFIX commands:

FIX(@IDESCENDENTS("Calendar Periods"))

"Gross Stock" = "Production" – "Sales";

ENDFIX

Notice how the FIX command has an area to list selection dimensions Within the

parenthesis you can list multiple dimension names separated by commas, member

names, or a list of member functions that returns member names There are also

functions that return members at specific levels in the outline You can even use basic structuring similar to a mathematical formula with AND/OR statements bracketed by extra parenthesis

In the example immediately above, notice how we are calculating the value for

Gross Stock for all Calendar Periods If your database was extremely large, with several years of historical data in it, this would be inefficient to always calculate the

entire Time dimension.

Now, suppose you added User Defined Attributes (UDAs) to the periods in the time

dimension and only wanted to calculate current periods? Here is what you would do:

1 Add UDAs of Historical, Current, and Forecast to the individual year

members and their children in the database outline For example, if the

current calendar year is 2009, then 2009 and all of its children would have

the UDA of Current added to them in the outline All years and their

children prior to 2009 would get the UDA of Historical All years and their

children later than 2009 would get the UDA of Forecast added in the outline:

2 Decide that you only want to calculate GrossStock for the Current

and Forecast time periods

3 Recode your FIX and ENDFIX command statements to look like this:

FIX(@UDA("Calendar Periods","Current") OR

@UDA("Calendar Periods","Forecast"))

"Gross Stock" = "Production" – "Sales";

ENDFIX

Aren't FIX and ENDFIX commands the greatest?

Always remember with FIX and ENDFIX, you want to Fix on Sparse dimensions and IF on Dense dimensions

EXCLUDE/ENDEXCLUDE

Well there you have it! With the release of Essbase version 9.x we now have EXCLUDE

and ENDEXCLUDE commands to compliment the FIX and ENDFIX commands

Now, without confusing you, we can tell you to consider the FIX and ENDFIX

commands as an include statement that tells Essbase what data to keep for calculating And, if including data proves difficult (and there are times it will), then you now have the EXCLUDE and ENDEXCLUDE commands to tell Essbase what data to leave out when

selecting data for calculating See examples in the next section

When we talked earlier about Essbase database calculation control, we meant the

FIX and ENDFIX commands Well, we also mean the EXCLUDE and ENDEXCLUDE

commands These commands complete the data control gate keeper job in the

Essbase database calculation script

The EXCLUDE and ENDEXCLUDE command is only for use in an Essbase database calculation script When you code a member formula, the formula applies only to the member in which it is coded, there is no need for EXCLUDE and ENDEXCLUDE commands

Because these database calculation commands are coded exactly like the FIX and

ENDFIX commands, we do not believe that we need to spend a great deal of time

explaining what should be obvious as far as the syntax is concerned

Remember the following points while using the EXCLUDE/ENDEXCLUDE command:

• You can nest EXCLUDE and ENDEXCLUDE command statements However,

make sure you only code arguments for a dimension in one EXCLUDE

statement no matter how many nested levels you have similar to the

FIX command

• You can code many instances of EXCLUDE and ENDEXCLUDE commands in

the same calc script

• You must have an ENDEXCLUDE for every occurrence of the EXCLUDE command

• You do not need to end an EXCLUDE or ENDEXCLUDE statement with a

semi-colon

• You code a comma separated list of members or functions that return a list

of members between the parentheses on the EXCLUDE command

• The EXCLUDE statement tells Essbase what data to leave out of the calculation

Basic EXCLUDE/ENDEXCLUDE examples

To begin with, as mentioned earlier, EXCLUDE and ENDEXCLUDE commands

can be nested We have also given you tips regarding when to use the EXCLUDE

and ENDEXCLUDE commands and what to avoid when nesting them, like the

same dimensions named in separate EXCLUDE commands

For starters, this is how you would code the simplest of EXCLUDE/ENDEXCLUDE

commands:

EXCLUDE("Canada")

"Gross Stock" = "Production" – "Sales";

ENDEXCLUDE

Notice how the EXCLUDE command has an area to list dimensions Within the

parenthesis you can list dimension names separated by commas, member names,

or a list of member functions that return member names There are also functions

that return members at specific levels in the outline You can even use basic

structuring similar to a mathematical formula with AND/OR statements bracketed by extra parenthesis

In the example immediately above, notice how we are calculating the value for

Gross Stock for all markets except Canada

There has not been a tremendous amount of feedback regarding the

use and performance of EXCLUDE and ENDEXCLUDE as they are new in

version 9.x We can only assume that the EXCLUDEandENDEXCLUDE

commands function exactly like the FIX and ENDFIX commands, except

to exclude the referenced data instead of including the referenced data

Aren't EXCLUDE and ENDEXCLUDE commands the greatest as well?

Functional

In Essbase database calculation scripts, you may be required to perform certain

functional commands like Set functions, Cleardata functions, or the Datacopy

function that is used to copy data from one member to another member, all before

you start the execution of your database calculation

SET command functions

SET command functions are used to tune the calculation and database factors

that affect calculation performance and are initially coded at the beginning of a

calculation script SET commands stay in effect for the duration of the calc script

unless coded again mid-script

Let's take a look at some of the widely used SET command functions:

• SET AGGMISSING ON/OFF: This command specifies if Essbase should turn

the consolidation of the # missing values ON or OFF This is only applicable

to Sparse dimensions

Example:

SET AGGMISSING ON;

SET AGGMISSING OFF;

• SET CACHE HIGH/DEFAULT/LOW/OFF/ALL: In the Essbase configuration file,

if you have specified values for the CALCCACHELOW, CALCCACHEHIGH, and

CALCCACHEDEFAULT, then when you run a database calculation script you

set the calculator memory cache to HIGH, LOW, or DEFAULT When you preset these values in the script (as needed), you can realize a better performance

while the script is executing

The Essbase configuration file is stored in the Essbase program directory

on the server This file can contain many entries for tweaks and adjustments that help your Essbase system perform better The Essbase Configuration

file is explained in more detail in Chapter 9 of this book

Let us say the following are the settings for the calc cache in the Essbase

configuration file Oops! We introduced a new term here The Essbase

configuration file is new and we have not talked about it yet Well, we

will discuss the configuration file in the coming chapters However, for

now know that it is a file where you can set up your server settings and

adjustments which improve the overall performance of data loads

and data calcs

Now, where were we? Oh yes, the server calc cache settings:

CALCCACHEHIGH 150000000

CALCCACHEDEFAULT 100000000

CALCCACHELOW 50000000

In the Calculation Script you have defined:

SET CACHE HIGH;

This command sets the calculator cache to 150,000,000 bytes

SET CACHE DEFAULT;

This command sets the calculator cache to 100,000,000 bytes

SET CACHE LOW;

This command sets the calculator cache to 50,000,000 bytes

• SET CALCTASKDIMS n: Specifies the number of Sparse dimensions that can run in parallel, where n is maximum number of Sparse dimensions This

command can help speed up the performance of a calculation script by

identifying tasks or transactions that can be run in parallel instead of serially

SET CALCTASKDIMS 1

In the above command, the last Sparse dimension is identified as the task

that can run in parallel

• SET CALCPARALLEL n: Enables the calculation tasks to run in parallel, where

n is number of threads The value of n can be 1 to 4 for 32 bit and 1 to 8 for 64 bit servers

SET CALCPARALLEL 4