Unique Essbase Customizations Using Java – Custom Logging

Introduction

Many companies have in depth working knowledge of Hyperion Essbase and are looking to enhance their enterprise reporting capabilities to the next level. Companies typically have specific processes and calculations that set them apart in their industry. However, they are often limited to basic reporting capabilities provided by the standard functions in Essbase. Additionally, complex operations can quickly become arduous using Calculation Scripts and Business Rules. This post will demonstrate the how to easily build Custom Java Routines to extend Essbase and dramatically reduce development time.

Complete details will be provided on how to implement a simple customized logging function for use in Calculation Scripts and Business Rules. Essbase’s streamlined, parallel nature makes it difficult for application developers to trace line by line. By using Java to implement a custom logging routine, one may use personalized log entries within their Essbase scripts. Consequently, developers can add tracing to their scripts and quickly determine how Essbase is approaching each calculation. Accordingly, application developers are able to see exactly how the script is being executed – providing quick debugging and faster development time. One powerful feature is to help determine block creation  within FIX statements.

The first step to integrating a custom Java routine into Essbase is to write some simple Java code. It is very easy – the code does not have to include any special APIs for Essbase.  During development, a few issues were encountered where Essbase was a bit picky about how the code is written.  Here are a few tips to help in getting started. These tips were gathered while doing real development, and it is best to follow at first, though you may revisit the items and find out what will work for you.

  • Do not include the code in a package such as “com.company.product_name” – remove the “package” declarative at the top of the code
  • Do not use the keyword “this” to refer to variables
  • Do not overload methods
  • Set all methods and variables to static

With these provisions in mind, the following code can be written to implement a custom logging routine.

CustomLoggerV2.java

import java.io.FileWriter;
import java.io.Writer;
import java.io.PrintWriter;
import java.io.StringWriter;
import java.util.Calendar;
import java.util.ArrayList;

public class CustomLoggerV2
{

    private static String logFile;
   
    public static int logFilterLevel;
   
    public static void setLogFilterLevel(int logFilterLevel2)
    {
        logFilterLevel = logFilterLevel2;
    }
   
    public static void setLogFilename(String logFilename)
    {
        logFilterLevel = 0;
        logFile = logFilename ;
    }

    public static synchronized void customLog (int logLevel, String message)
    {
        log(logLevel, message);
    }

    private static synchronized void log (int logLevel, String message)
    {
       
       
        // do not log
        if (logLevel < logFilterLevel)
            {
                return ;
            }
       
        try {
       
            Calendar c = Calendar.getInstance();
           
            FileWriter fw = new FileWriter(logFile, true);
               
            fw.write(c.getTime()   ": "   message   "\n");
            fw.close();
             
        } catch (Exception e)
            {
                System.out.println("Error, cannot open , "   logFile);
                e.printStackTrace();
            }
        }

 
}

The code implements three public methods:

  • setLogFilterLevel(int logFilterLevel) – sets the minimum message level to log (think about ERROR=100, WARN=90, INFO=70, DEBUG=0) – so you can easily change the verbosity of the output.
  • setLogFilename(String filename) – The full path to the log file you wish to use
  • customLog(logLevel, String message) – The log message, with its indicated priority

The next step is to package up the code above. It is important to use the same version of Java which is running your Essbase instance. To find the version, look for the JRE being used within the environment, for instance, Hyperion\common\JRE\Sun\1.5.0\bin. To obtain the specific revision, open a cmd prompt, cd to the bin directory, and run “java –version”.

E:\Hyperion\common\JRE\Sun\1.5.0\bin>java -versi
java version "1.5.0_11"

Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_11-b03)

Java HotSpot(TM) Client VM (build 1.5.0_11-b03, mixed mode

To compile the code a JDK is required, which will contain the javac command. Hyperion only packages the JRE, meaning you will have to download the correct JDK in order to compile the code. You can find older versions of Java JDK from Oracle(Sun)’s web site. Once you have obtained the correct version of the JDK, compile and package up the code:

javac CustomLoggerV2.java

jar -cf CustomLoggerV2.jar CustomLoggerV2.class

Next, copy the CustomLoggerV2.jar file into the Essbase file structure:

Copy CustomLoggerV2.jar into the E:\Hyperion\products\Essbase\EssbaseServer\java\udf folder. If the udf folder does not already exist, create it.

Now it is time to start including the Java class within Essbase. Essbase runs within its own JVM and therefore has its own Java security. In the example above, we are writing to a local log file, which will violate the default security policy setup in the udf.policy file. The file is usually found in Hyperion\products\Essbase\EssbaseServer\java . The simplest way to get around the security concerns for development purposes is to remove the comment from the last line in the file, which effectively includes the directive “permission java.security.AllPermission”

permission java.util.PropertyPermission “java.vm.version”, “read”;

permission java.util.PropertyPermission “java.vm.vendor”, “read”;

permission java.util.PropertyPermission “java.vm.name”, “read”;

// Uncomment the following line if you want to remove all restrictions

permission java.security.AllPermission;

};

Now that the Essbase security and jar file have been put in, a restart of the Essbase process is required to register the changes. Please restart Essbase now.

The final step is to run some maxl statements to register the public java methods with Essbase.

CustomLoggerV2.mxl

create or replace function '@JCustomLoggerV2_setLogFilename'

as 'CustomLoggerV2.setLogFilename(String)'

spec '@JCustomLoggerV2.setLogFilename(absolute file name)'

comment 'Nicholas King'

with property runtime;

create or replace function '@JCustomLoggerV2_customLog'

as 'CustomLoggerV2.customLog(int, String)'

spec '@JCustomLoggerV2.customLog(log level, log message)'

comment 'Nicholas King'

with property runtime;


create or replace function '@JCustomLoggerV2_setLogFilterLevel'

as 'CustomLoggerV2.setLogFilterLevel(int)'

spec '@JCustomLoggerV2.setLogFilterLevel(filter level)'

comment 'Nicholas King'

with property runtime;

One final thing… In order to run a custom java function, the value of the result has to be stored in an Essbase member. This is true even if there is not any use for the return value, such as this case where there is no value returned from the Java methods. To get around this, create a new Essbase member called “No Measure” somewhere within your Essbase outline. This will act as a dummy member intended only to direct the return value, if any, of the Java methods. An example is shown below.

Sample Calc Script or Business Rule to Invoke the Logger

//ESS_LOCALE English_UnitedStates.Latin1@Binary

/* SETUP The Logger */

/* Fix on something so it runs only once */

FIX (Actual, Texas, "100-10")

"No Measure" = @JCustomLoggerV2_setLogFilename("E:\CustomEssbaseLog.log");

"No Measure" = @JCustomLoggerV2_setLogFilterLevel(50);

ENDFIX;

/* In your script, do some actual logging */

FIX (Actual, Texas, "100-10")

/* Won’t be displayed */

"No Measure" = @JCustomLoggerV2_customLog(0, "This is a debug message");

/* Will be displayed */

"No Measure" = @JCustomLoggerV2_customLog(50, "This is a normal message");

"No Measure" = @JCustomLoggerV2_customLog(100, "This is an important message");

ENDFIX;

The result of running the script is the log entries will be added to the log file E:\CustomEssbaseLog.log,

Mon Feb 21 01:30:25 EST 2011: This is a normal message

Mon Feb 21 01:30:25 EST 2011: This is an important message

Troubleshooting Tips

A very common error you may receive is,

Error: 1200324 Error compiling formula for [No Measure] (line 8): operator expected after [@JCustomLogger_customLog]

This error is a generic error that indicates something in your custom function is not registered properly.  Unfortunately, there is not a lot of detailed log information at this point to help discover the problem. If you receive this message a few things might help:

  • Retrace your steps – carefully review all instructions above
  • Check that the correct version of Java was used to compile the class file and package the jar
  • Check the jar is in the correct “udf” folder in Essbase
  • Check the syntax of the MAXL to register the functions is correct
  • Simplify your script as much as possible to reduce the possibility of syntax errors

Conclusion

This example shows how to create a custom Java based logger integrated into Essbase. The possibilities are endless – anything that can be done in Java can be added to Essbase. You can create development aids, or even read/modify the values within the cube. For instance, this model has successfully been used to perform complex financial calculations within Hyperion Planning Forms using Business Rules.  It could also be used for integrating Web Services with your cube by reading or writing cube data and interacting with an enterprise Web Service.




Initial Release of In2Hyp Services Manager – Restarting Services Just Got Easier!

As Hyperion applications have become more integrated, the need for multiple servers to support the environment is now required.  Although the recent releases of version 11 are less dependent on the order in which the services start, Hyperion still recommends a specific order to start the services so they perform correctly.  As these services are typically on multiple servers, it is time consuming to perform this operation and it is prone to error.  Many organizations lack the ability to automate this task when services are interrupted, or operating systems are updated, and there is no automated way to start and stop the services.

In2Hyperion is now making available HypServicesManager.  HypServicesManager is an application that will start or stop the services in a predefined order.  An XML file holds the server/service order so the services to be included, and the order in which they start/stop, is completely customizable.  It will impersonate a domain account that has permissions to the respective servers, so the starting or stopping of the services on multiple servers can be performed quickly and remotely.  With command line parameters, it can even be automated.

Keep in mind, HypServicesManager is completely independent of Hyperion services.  Although it was developed to fill the need to manage Hyperion services, it can be used for any windows service as it uses the windows APIs.

I have used this application in multiple organizations and it is often used in a production environment.  With that said, this has not been tested on thousands of environments in every Windows OS.  If it doesn’t work for you, we certainly welcome feedback and will make every effort to fix any bugs that are discovered.

Details on the use of the application are available here.




Empower Users To Improve Calculation Times

As an Essbase user, you have more power to improve performance than you think.  How many times do you lock and send data through Excel, SmartView, or web forms, that include zeros?  How many times do you allocate data to a finite level out of convenience?  Understanding what this does to Essbase is critical to understanding how a user can negatively impact performance without adding any value to the analysis or the results the database produces.

I analyzed a planning database used in one of the largest financial institutions in the world.  Over 60% of the values entered were zero.  Another 20% of the values were less than 1 dollar.  By eliminating the zeros, the total calculation time of the planning application was under 20 minutes.  With the zeros, it was nearly 2 hours.

There are two reasons for this.  First, there is a different between empty and zero.  Empty consumes no space to store whereas a zero consumes the same space as 1 billion.  Think of this as a grocery bag.  If you fill a grocery bag with nothing, it takes up no space.  If you fill it with empty cans (a zero), it consumes the same amount of space as if those cans were full (1 billion).

The example below is very common.  Assume that a forecast needs to be done for the last 3 months of the year.  Frequently, a spreadsheet would hold zeros for the first 9 months.  18 cells have zero and 6 cells have a positive value.  That means that 75% of your data could be eliminated by not loading zeros.

The same load with #Missing is more effective.

I highly recommend reading the article explaining dense and sparse to understand what a block is and what it represents before you continue this article.

There is also another very significant factor in loading zeros.  Loading a zero that creates a block just to hold a value of zero can explode the size of the database, as well as the time it takes to consolidate and execute business rules.  The more blocks that have to be loaded and consolidated, the longer it takes to finish.  If each block was a spreadsheet and you had to do this manually, you would have to open each spreadsheet and enter the number into a calculator to consolidate.  If 75% of the blocks you opened were zero, it wouldn’t change your total, but it would drastically increase the time it takes because you still have to open each spreadsheet.   If an Essbase database has 1,000 blocks, and 75% of them only hold zeros, it will likely take 2 or 3 times longer to calc the zeros because it still has to open the block and add the zero.  Remember, a zero acts no differently than a value of 100.

As an example to the above, the following example would create a block for South and West, inflating the database size.

 

Users can significantly reduce this unnecessary explosion in size by loading a blank as apposed to a zero.  If zeros are already in the database, leaving the cell blank will NOT overwrite the zeros.  If zeros are loaded inadvertently, a #Missing has to be used to remove them.

For all you users loading data, it can be a hassle removing the zeros.  Being responsible can significantly improve your experience with Essbase.  To make it easier, take a look at the function in the In2Hyperion Excel Ribbon that replaces all zeros with #Missing.




Altering Large Numbers Of Cells In Excel A Hundred Times Quicker

Many processes need to write large volumes of data in Excel.  The typical method is to loop through each cell and perform the action.

   Dim CellsDown As Long
   CellsAcross As Long

   Dim CurrRow As Long
   CurrCol As Long
   Dim CurrVal As Long

   '  This can be replaced with the selected range and is just used to illustrate this example.
   CellsDown = 1000
   CellsAcross = 36

   '   Loop through cells and insert values
   CurrVal = 1
   Application.ScreenUpdating = False
   For CurrRow = 1 To CellsDown
       For CurrCol = 1 To CellsAcross
           Range("A1").Offset(CurrRow - 1, CurrCol - 1).Value = CurrVal
           CurrVal = CurrVal   1
       Next CurrCol
   Next CurrRow

Rather than writing the values out cell by cell, it is quicker to store the value in an array and write the array to a range of cells at one time.

   Dim CellsDown As Long
   CellsAcross As Long

   Dim CurrRow As Long
   CurrCol As Long
   Dim CurrVal As Long

   Dim TempArray() As Double

   '  This can be replaced with the selected range and is just used to illustrate this example.
   CellsDown = 1000
   CellsAcross = 36

   '   Update the array
   ReDim TempArray(1 To CellsDown, 1 To CellsAcross)
   Set TheRange = Range(Cells(1, 1), Cells(CellsDown, CellsAcross))

   '   Fill the temporary array
   CurrVal = 0
   Application.ScreenUpdating = False
   For i = 1 To CellsDown
       For j = 1 To CellsAcross
           TempArray(i, j) = CurrVal
           CurrVal = CurrVal   1
       Next j
   Next i

   '   Transfer temporary array to worksheet
   TheRange.Value = TempArray

This same method can be used when altering data.  By changing the following line

            TempArray(i, j) = CurrVal

To this

            TempArray(i, j) = TheRange(i, j) * 3

By using TheRange(i, j), the existing value can be altered

 

The process of writing values cell by cell took 3.16 seconds.  Using the array method, it took .08 seconds, nearly 40 times faster




Getting large numbers in FR Reports?

When using Workspace to view reports, some users have seen excessively large numbers that don’t belong. If you are having this issue, it could very well be because the default Essbase query engine in 11.1.1.x is the MDX query engine, which can cause documented bug 9062413. Essentially, this bug will cause users to see the same astronomical number in every cell that sits on an intersection to which the user does not have security access. Understandably, this can cause some concern. This issue is expected to be fixed in a future release, but until then, the query engine will need to be manually changed.
The first option is to fix the issue at a report level. This is a relatively quick process and is a good idea if you only have a handful of reports.  To change the query engine setting for a particular report, follow the steps below:

1. Open the report.
2. For each grid, select the entire grid.
3. Right-click and select Data Query Optimization Settings.
4. Deselect the option “Essbase Queries Use MDX.”
5. Save the report.

Repeat the above process for each report.

For users that have a larger number of reports, a better option may be to change the query engine in the properties file on the server.  The benefit to fixing the issue in the properties file is that changes only need to be made once, and all reports will reflect this change.
The file that needs to be edited is located on the Financial Reporting (app) server, typically on the path D:\<Hyperion Home>\products\biplus\lib where <Hyperion Home> represents the root location of the Hyperion install. The file that needs adjusted is the fr_global.properties file.
Open the properties file and add these lines:

# MDX Query Engine has been set as the default in Essbase 11.1.1.x. This can cause bug 9062413 
# which may cause unauthorized users to see a long series of numbers in each cell when running  
# reports. To solve this issue, the below line was added, which switches the query engine.
EssbaseUseMDX=false

Any line preceded by “#” is commented out. Therefore, these can say whatever you prefer, but should give anyone that views this file a good indication why this text is in the file.

Once the properties file has been updated, the following services must all be stopped in the following order, then started in the same order for the changes to take effect.

1. Hyperion Financial Reporting – Print Server
2. Hyperion Financial Reporting – Report Server
3. Hyperion Financial Reporting – Scheduler Server
4. Hyperion Financial Reporting – Web Application (Note – This service may be on the FR (Web) server, not the FR(App) server like the other three services.)

Note – This modification will apply to everyone using the server on which they are made, so be careful when making changes to a shared server.




Altering Large Numbers Of Cells In Excel A hundred Times Quicker

Many processes need to write large volumes of data in Excel.  The typical method is to loop through each cell and perform the action.

   Dim CellsDown As Long
   CellsAcross As Long

   Dim CurrRow As Long
   CurrCol As Long
   Dim CurrVal As Long

   '  This can be replaced with the selected range and is just used to illustrate this example.
   CellsDown = 1000
   CellsAcross = 36

   '   Loop through cells and insert values
   CurrVal = 1
   Application.ScreenUpdating = False
   For CurrRow = 1 To CellsDown
       For CurrCol = 1 To CellsAcross
           Range("A1").Offset(CurrRow - 1, CurrCol - 1).Value = CurrVal
           CurrVal = CurrVal   1
       Next CurrCol
   Next CurrRow

Rather than writing the values out cell by cell, it is quicker to store the value in an array and write the array to a range of cells at one time.

   Dim CellsDown As Long
   CellsAcross As Long

   Dim CurrRow As Long
   CurrCol As Long
   Dim CurrVal As Long

   Dim TempArray() As Double

   '  This can be replaced with the selected range and is just used to illustrate this example.
   CellsDown = 1000
   CellsAcross = 36

   '   Update the array
   ReDim TempArray(1 To CellsDown, 1 To CellsAcross)
   Set TheRange = Range(Cells(1, 1), Cells(CellsDown, CellsAcross))

   '   Fill the temporary array
   CurrVal = 0
   Application.ScreenUpdating = False
   For i = 1 To CellsDown
       For j = 1 To CellsAcross
           TempArray(i, j) = CurrVal
           CurrVal = CurrVal   1
       Next j
   Next i

   '   Transfer temporary array to worksheet
   TheRange.Value = TempArray

This same method can be used when altering data.  By changing the following line

            TempArray(i, j) = CurrVal

To this

            TempArray(i, j) = TheRange(i, j) * 3

By using TheRange(i, j), the existing value can be altered

 

The process of writing values cell by cell took 3.16 seconds.  Using the array method, it took .08 seconds, nearly 40 times faster