Adventures in Groovy – Part 20: Groovy On-Premise vs. Groovy Cloud

Introduction

Yes, it is true that Groovy is available in on-premise and cloud (PBCS) versions of Hyperion Planning.  No, it is not true that the same flavor of Groovy exists in both.  Both have their advantages, and both have their drawbacks.  The likelihood that they will ever be the same is extremely low, and here is why.

The Difference Is

On-Premise gives developers the ability to write and use independent Groovy compiled applications.  These can be used in Business Rules as CDFs (custom defined functions).  Developers have complete functionality to make this do whatever they want.  It can return results to save to Essbase/Planning, it can interact with SQL, can run other programs, pretty much anything you can access that has a JAVA API.

PBCS doesn’t have the same flexibility.  Custom defined functions can’t be compiled and stored on the server.  PBCS, rather, has “Groovy Calculations.”  This gives developers the flexibility to interact with the Data Forms that on-premise doesn’t have.  Developers can iterate through the cells and act accordingly.  It can stop the form from saving, calculate and override data entered, color code cells, customize Data Maps, Smart Pushes, dynamically generate calculations, move data between databases, all with access to much of the Groovy functionality.

PBCS also supports the REST API, so Groovy can be used to access that and do everything, even more, that EPM Automate can do.

Why They Will Never Be The Same

This is just an opinion.  Technology changes so rapidly that this may change.  Corporate strategy changes almost as rapidly.

If PBCS had to ability to do what on-premise does, the ability for Oracle to support the instance would be a challenge.  CDFs can delete all the files on a server, for instance, and I don’t see a cloud provider giving developers this much control in a shared environment.

I also don’t see on-premise to have the same proactive interaction that PBCS has with Groovy Calculations purely because Oracle is pushing the cloud, and they want the most current functionality to exist in the platform they are pushing clients to use.

My Two Cents

I understand why there is a difference, and I don’t expect it to change in the near future.  3 years ago I didn’t expect that I would tell you that I would rather do a cloud implementation than on prem, either.  I do think as people get more comfortable with the cloud, and security improves, there will be advances.  I think there will be a future state where the cloud offerings will be closer to having the flexibility to the on-premise implementations.

 




Adventures in Groovy – Part 19: Real Time Reporting Webinar with Breakthru Beverage Group

Introduction

Chris Hull has been kind enough to partner with us to present how the methods available in Groovy calculations have made a huge impact in their budgeting and reporting process using PBCS.

One of the biggest user complaints about their first budgeting process in PBCS was that they had to wait to get their consolidated reporting until an admin ran the process or a scheduled consolidation ran.  Thanks to what Groovy offers, this is no longer going an issue.

For those of you that participated, I have added some additional content.  I know we were limited on time.  I skipped a few examples and rushed through diving into the actual steps and Groovy processes.  I included additional examples and more information on the process below.  If you didn’t attend, I would encourage you to walk through the items in the order they are presented.

If any of this is of interest to you and you would like to speak further about the capabilities, please send an email.

Thanks for everybody that participated and supported this event.

Presentation Recording

Additional Questions

There were a few questions after we signed off.  I will do my best to answer them below.

Question: How can internal IT team support this solution? Does Huron train them on how to to write Groovy, change code in Groovy?  Demo shows only 3 cells were changed, how is performance if all products were changed? Doesn’t it hit MDX error limit anyways?  Does Oracle SR team support issues/bugs related to Groovy?

Answer: Most of our solutions are not maintained by IT, but by finance.  Huron certainly can train administrators on how to maintain the Groovy calculations.  As far as performance, I changed 75 products and it took 14 seconds to push the data from GP to the GP reporting application, and .4 seconds to synchronize it to the Fin cubes.  MDX isn’t used to do the synchronization so the MDX text limit is irrelevant.  Oracle does support the product and they support issues related to the API, which Groovy uses.  I have seen some issues resolved in weeks, and others that are still being addressed months later.

Question: What capabilities does the tool have to track changes to the budget?

Answer: PBCS, when turned on, has auditing at the cell level.  Every time data is changed, the user, date/time, old value, and new value are recorded.  At any point a user can look at that and see the change history.

Expanding On The Groovy Form Save Process



Additional Examples

Presentation PowerPoint



Conclusion

Thanks again.  I hope to do more of these in the future and will make sure  you are included in the invitation.




Adventures in Groovy – Part 18: Real Time Data Movement (Setting The Stage)

Introduction

One of the challenges with Hyperion Planning is the ability to move data between applications in real time.  A classic example of this is a P&L application with other modules that have greater detail.  The following is an example.

  • A Gross Profit specific database that includes a product, delivery channel, and product type dimension.
  • A CapEx specific database with asset type, asset, and asset category
  • A Workforce specific database with job type, union, and employee.
  • A P&L application that includes income and expense with information fed from the detailed models at consolidated levels.

In June of 17, with the release of Groovy Calculations, the ability to update any of the detailed models and synchronize the consolidated data in real time to the P&L database became possible.  When a user saves data, within seconds, the data can be reflected in a database with different dimensional.

Setting The Stage

This is going to be a lengthy, multi part article.  Before we begin, the application architecture is going to be laid out so the calculations can be explained in detail.  The application will consist of 2 play types.  The first is the P&L and the second is a detailed product planning play type.  We won’t introduce a Capex and Workforce model.  It will only complicate the explanation and is redundant in the logic required.

The data flow and architecture looks like this.GP (Gross Profit Product Detail) databases

The initial plan type is called GP

Although this may not match with your model, the concept is the same.

  • It has dimensions that are required to plan at a product level that don’t exist in the P&L application.
  • It has specific logic that doesn’t apply to other databases.
  • It has a unique account dimension that doesn’t mirror what is in the other applications.
  • Consolidation takes a long time and is not optimal to be performed on a data form save.

As previously stated, the same differences will exist in other models, like Capex and Workforce.

Fin (Income Statement / Balance Sheet) databases

The Fin application is a typical consolidated reporting application that excludes details like product level revenue, employee level plans, and assets and their properties needed to calculate capital expense.

Dimensional Summary

For this example, the following shows the application dimensions and database associations

The Synchronization Process

The GP database includes 3 dimensions that don’t exist in the Fin model.  For this to be moved to the Fin model, 3 dimensions need to be consolidated.  The GP model also has a different account structure.  A translation between the two account structures has to occur before the synchronization can be completed.  The other piece that is not required, but highly encouraged, is to only work with the data that has changed.  So, this will dynamically select the data rows on the form that have been edited by the user.  Functionally, the following happens when a user saves a data form.

  • Identify the members that need to be included in the synchronization
  • Push the level zero data from the GP BSO database to the GP ASO database (only edited data)
  • Retrieve the data from the GP ASO database at a total product, channel, and material group
  • Submit the data from the above retrieve to the Fin BSO application and the rFin ASO application
  • Execute any logic that needs to be completed in the Fin application (taxes, driver-based data, etc.)
  • Push the level zero data from the Fin BSO database to the Fin ASO database

Groovy Methods Required

There is a lot going on here, so we are going to summarize and explain the Groovy methods that will be used to accomplish the synchronization.

DataGridIterator

To make this as efficient as possible, it is important to only execute the methods on the data that have been edited.  If you haven’t read Part 3 of this series, take a look before you continue.

DataMap / SmartPush

Once the POV is identified that needs to be included in the synchronization, the first operation is to push that data to the reporting cube.  This will be used a couple of times in this sequence.  Part 8 of the Groovy Series covers this in detail and an understanding is helpful before you continue.

DataGridBuilder / DataGridDefinitionBuilder

This has not been covered yet.  These methods give you complete control to simulate a retrieve and submit.  These two objects are the major pieces of the puzzle that have never really been exposed in any fashion.  These are the methods that really open up the possibilities for real time reporting.

Take A Breath

You may be a little overloaded with new information.  We will let this settle in and give you a chance to digest the concepts.  The next article will walk you through the code.  To satisfy your curiosity, watch this video, which takes you through the above example in a live environment.




Will Groovy Calculations in PBCS Solve The World’s Problems?

No, But Can It Solve Yours?

I received a lot of positive feedback on the Groovy Series and have been asked a many great questions.  People are excited about the improvements but are still a little hesitant to buy in to the hype.  They question, and rightfully so, 

  1. are the performance gains really as positive as I have stated, and
  2. is the functionality that can be added to improve a user’s experience really available, and
  3. can it improve the validity of the data input as much as I have said it does?

Challenge Accepted!

If you have a challenge, performance issues, or missing functionality that you desperately need in Data Forms, post a quick comment with a summary about what you are facing.  Please enter a valid email so I can contact you directly with any questions.  Don’t worry, your email will remain confidential!

I am going to try to pick one situation every week or two and provide some alternatives with Groovy that will solve, or improve the problem you are facing.

I encourage you to

Join the Party!!function(m,a,i,l,s,t,e,r){m[s]=m[s]||(function(){t=a.createElement(i);r=a.getElementsByTagName(i)[0];t.async=1;t.src=l;r.parentNode.insertBefore(t,r);return !0}())}(window,document,'script','https://in2hyperion.com/wp-content/plugins/mailster/assets/js/button.min.js','MailsterSubscribe');

so you are notified about the solutions posted.

Good luck on your quest to make Hyperion Planning a better experience for all of your users!




Adventures in Groovy – Part 17: Force Cell Comments

Challenge Accepted

When I asked visitors to try to come up with a situation that Groovy Calculation might be able to solve, this was a good one.  One visitor asked if they could require a cell comment if certain parameters were not met.  It is actually relatively easy.

The following requirement exist in this example.  If any month holds more than 30% of the full year, that cell requires the user to enter a cell comment.  If no comment exists, the user won’t be able to save the form.

The User Experience

If any month is more than 30% of the full year, and the user doesn’t add a comment to a cell, the form will not save.  The following shows what happens when the above fails, and what happens after the user enters a comment into the cell.


The Code

def backErrColor = 16755370 //Red
def caseTotal = 0
def accountName = ""

// Loop through the months
operation.grid.dataCellIterator('Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec').each { 
  // Get a total for all 12 months every time the row changes
  if(it.getAccountName() != accountName) {
    accountName = it.getAccountName();
    caseTotal = it.data + it.crossDimCell('Feb').data + it.crossDimCell('Mar').data + it.crossDimCell('Apr').data + it.crossDimCell('May').data + it.crossDimCell('Jun').data + it.crossDimCell('Jul').data + it.crossDimCell('Aug').data + it.crossDimCell('Sep').data + it.crossDimCell('Oct').data + it.crossDimCell('Nov').data + it.crossDimCell('Dec').data 
  }
  // If the value is greater than 30% of the total and the cell does NOT have a cell comment, interrupt the form save
  if(it.data > 0 && it.data / caseTotal > 0.3 && !it.hasCellNote() ) {
    it.addValidationError(backErrColor, "Cases for a single month can't be more than 30% of the total year without an assumption.", false) 
  }
}

Conclusion

Challenge accepted.  This one goes in the win column for Groovy Calculations!




Adventures in Groovy – Part 16: Ignore Form Save When No Data Has Been Edited

Introduction

I know you can argue this is a user issue and a training issue, but the fact is, sometimes people will save a form without editing any data.  There are at least three negative issues as a result.  One, the business rules and smart pushes execute, consuming unnecessary resources.  Two, users may think they made changes and expect changes in the results.  Three, if the processes are time consuming (like applying allocations or currency rates globally), the user will have to wait to correct the issue.  There is a very simple way to stop all the processes from executing and inform the user that they haven’t made any changes.

The Code

The following will provide a template to use to accomplish the interruption of the form save.  The messageBundle can be altered to be whatever is required and in as many languages as needed.  If this is new to you, read Part 13 of this series.

//Setup Message Bundle
def mbUs = messageBundle( ["validation.NoDataChanged":"No data was altered so no business logic was executed."])
def mbl = messageBundleLoader(["en" : mbUs])

//Create the grid and iterator objects
DataGrid curgrid = operation.getGrid()
GridIterator itr = curgrid.getDataCellIterator(PredicateUtils.invokerPredicate("isEdited"))
// Throw an exception if no cells are edited
if(!curgrid.empty && !itr.hasNext() ){throwVetoException(mbl, "validation.NoDataChanged")}

A second option would be to count the cells that have been edited.  One reason to show this, as it is not my first choice, is that it could be altered for another purpose.  If the number of cells edited is required, this could be used to count them.

def iCount = 0 

//Setup Message Bundle 
def mbUs = messageBundle( ["validation.NoDataChanged":"No data was altered so no business logic was executed."])
def mbl = messageBundleLoader(["en" : mbUs]) 

//Iterate through the edited cells
operation.grid.dataCellIterator({DataCell cell -> cell.edited}).each { iCount +=1 }

// Throw an exception if no cells are edited
if(iCount == 0) { throwVetoException(mbl, "validation.NoDataChanged") }

Implementing On A Form

Once the code is save as a business rule, add it to any form and run it BEFORE SAVE.  This example assumes the code above is saved as GP – IsEdited.

As long as it is set to run before save and there are no cells on the form that have been altered, the user will get an error that displays whatever the message bundle represents.

Summary

This isn’t going to change your life, or make a drastic improvement for your users.  But, it is a simple thing that is easy to implement that will add some polish to your application.  If you have found other ways to use Groovy Calculations to add some polish to your application, please share with a comment.




Adventures in Groovy – Part 15: Returning Errors (RTP Edition)

Introduction

One of the huge benefits that available in Groovy Calculations is the ability to interact with a user, validate data, and act on the validation.  Now, we can interrupt form saves, stop Run Time Prompts from continuing, and communicate information back to the user.

This may sound repetitive if you have read part 13 and part 14, and you can skip to the code example to learn more about run time prompt validation.  If not, you must have an understanding of the validation functions and the components of the messageBundle.There are a number of functions for validation, and they can be categorized functionally. Although they all can be use somewhat interchangeably, the logical uses are

  • Data Form validation functions
    • addValidationError
  • RTP validation functions
    • validateRtp
  • Validation functions that are more open ended and can be used just about anywhere
    • messageBundle
    • messageBundleLoader
    • throwVetoException

In this post, we will discuss one aspect of this, and probably the simplest application, validating Run Time Prompts (RTP).

The MessageBundle

Before a few of the methods can be used, one must first understand the MessageBundle and MessageBundleLoader methods.  To look at documentation, they might seem very complex, and a maybe a little intimidating.  The reality is that these are just objects that hold the error messages.  That is pretty much the long of short of it.  The messageBundle holds a map (basically a lookup table that is two columns and n rows) of the error ID and the description of the error you want to display.  If the application is consumed by users with multiple languages, a messageBundle can be created for each language.  The messageBundleLoader allows you to identify which bundle to use based on the user’s local.  The example below should answer any questions you have.

The Message Bundle

Think of this method as an array, or a table in Excel.  It has 2 columns (ID and message).  It can have an infinite amount of rows.  The syntax of this is “[id,message]”.  For multiple errors, this is duplicated, separated by a comma, like “[id,message],[id,message]”.  Here is an example of a messageBundle with one error.

def mbUs = messageBundle( ["validation.InvalidCharacters":"Only alphanumeric characters can be entered (a-z, 1-9)."] )

And with two errors.

def mbUs = messageBundle( ["validation.InvalidCharacters":"Only alphanumeric characters can be entered (a-z, 1-9)."],
["validation.Negative":"A positive number is required."])

And with two errors in Spanish.

def mbSpanish = messageBundle( ["validation.InvalidCharacters":"Sólo se pueden introducir caracteres alfanuméricos (a-z, 1-9)."],
["validation.Negative":"Se requiere un número positivo."])

This can be extended to hold all the error messages required for the scope of the calculation.

The Message Bundle Loader

The messageBundleLoader is the piece that pulls either a single, or multiple, messageBundles together to use in a call.  If only one language is required, it would look like this.

[def mbl = messageBundleLoader(["en" : mbUs])

For multiple languages, it would include multiple messageBundles

[def mbl = messageBundleLoader(["en" : mbUs],["en" : mbSpanish])

Validate The Input

When a validation error exists, the prompt window will not close, so it won’t let a user continue unless all the data entered validates.  Validations are only limited to your knowledge of how to validate the input.  Let Google be your friend.  You will be hard pressed to have a sitiation where you can’t find an example of what you are trying to do.  If you aren’t familiar with “regex,” it will likely be included in just about any Google search you do.  The examples below all use a regex string to validate the inputs.

To use a run time prompt in Groovy, they must be initiated.  This looks like a comment, but it acts differently when prefaced by RTPS:

/*RTPS: {EmployeeName} {EmployeePhone} {EmployeeEmail} */

Next, we will create a messageBundle.  Although it is simplier than above, it is more than enough to demonstrate its use in the validateRtp method.  This creates an error for each of the three validations in English.

def mbUs = messageBundle(["validation.invalidemail":"Email address is invalid: {0}", "validation.invalidphone":"Phone number is invalid: {0}", "validation.invalidnamelength":"Employee name must be 5 to 40 characters: {0}"]) 
def mbl = messageBundleLoader(["en" : mbUs])

Now, the actionable stuff.  The next 3 lines will validate the 3 run time prompts.  If any of them fail, the RTP window will remain open and the user can’t continue until they fix the errors or cancel the action.

// Validate the Rtp values
validateRtp(rtps.EmployeeName, {(5..40).contains(it.enteredValue.size()) }, mbl, "validation.invalidnamelength", rtps.EmployeeName)
validateRtp(rtps.EmployeeEmail, /^.+@.+/, mbl, "validation.invalidemail", rtps.EmployeeEmail.enteredValue)
validateRtp(rtps.EmployeePhone, /^(?:\+?1[- ]?)?\(?([0-9]{3})\)?[- ]?([0-9]{3})[- ]?([0-9]{4})$/, mbl, "validation.invalidphone", rtps.EmployeePhone)

Putting it all together, we have the following.

/*RTPS: {EmployeeName} {EmployeePhone} {EmployeeEmail} {Scenario} {Year} {Period} {Entity} {Version}*/
def mbUs = messageBundle(["validation.invalidemail":"Email address is invalid: {0}", "validation.invalidphone":"Phone number is invalid: {0}",
"validation.memberexists":"The member you have specified already exists and cannot be created: {0}.", "validation.invalidnamelength":"Employee name must be 5 to 40 characters: {0}"])
def mbl = messageBundleLoader(["en" : mbUs])

// Validate the Rtp values
validateRtp(rtps.EmployeeName, {(5..40).contains(it.enteredValue.size()) }, mbl, "validation.invalidnamelength", rtps.EmployeeName)
validateRtp(rtps.EmployeeEmail, /^.+@.+/, mbl, "validation.invalidemail", rtps.EmployeeEmail.enteredValue)
validateRtp(rtps.EmployeePhone, /^(?:\+?1[- ]?)?\(?([0-9]{3})\)?[- ]?([0-9]{3})[- ]?([0-9]{4})$/, mbl, "validation.invalidphone", rtps.EmployeePhone)

Wrap Up

It has been a long time since developers have had this kind of control.  The possibilities are only limited by your imagination and business requirements, but there isn’t any validation that can’t be done.  This wraps up the 3 validation methods.

Enjoy this new functionality.  Don’t underestimate its importance.  This functionality can save your customers hours of work and lots of frustration.  Helping them input accurate data improves the forecasting and budgeting process.  Implement these techniques and they will love you!

 

 




Adventures in Groovy – Part 13: Returning Errors (Data Forms)

Introduction

One of the huge benefits that available in Groovy Calculations is the ability to interact with a user, validate data, and act on the validation.  Now, we can interrupt form saves, stop Run Time Prompts from continuing, and communicate information back to the user.There are a number of functions for validation, and they can be categorized functionally. Although they all can be use somewhat interchangeably, the logical uses are

  • Data Form validation functions
    • addValidationError
  • RTP validation functions
    • validateRtp
  • Validation functions that are more open ended and can be used just about anywhere
    • messageBundle
    • messageBundleLoader
    • throwVetoException

In this post, we will discuss one aspect of this, and probably the simplest application, validating Run Time Prompts (RTP).

The MessageBundle

Before a few of the methods can be used, one must first understand the MessageBundle and MessageBundleLoader methods.  To look at documentation, they might seem very complex, and a maybe a little intimidating.  The reality is that these are just objects that hold the error messages.  That is pretty much the long of short of it.  The messageBundle holds a map (basically a lookup table that is two columns and n rows) of the error ID and the description of the error you want to display.  If the application is consumed by users with multiple languages, a messageBundle can be created for each language.  The messageBundleLoader allows you to identify which bundle to use based on the user’s local.  The example below should answer any questions you have.

The Message Bundle

Think of this method as an array, or a table in Excel.  It has 2 columns (ID and message).  It can have an infinite amount of rows.  The syntax of this is “[id:message]”.  For multiple errors, the id:message is duplicated, separated by a comma, like “[id,message,id:message]”.  Here is an example of a messageBundle with one error.

def mbUs = messageBundle( ["validation.InvalidCharacters":"Only alphanumeric characters can be entered (a-z, 1-9)."] )

And with two errors.

def mbUs = messageBundle( ["validation.InvalidCharacters":"Only alphanumeric characters can be entered (a-z, 1-9).",
"validation.Negative":"A positive number is required."])

And with two errors in Spanish.

def mbSpanish = messageBundle( ["validation.InvalidCharacters":"Sólo se pueden introducir caracteres alfanuméricos (a-z, 1-9)."],
["validation.Negative":"Se requiere un número positivo."])

This can be extended to hold all the error messages required for the scope of the calculation in all the locales required.

The Message Bundle Loader

The messageBundleLoader is the piece that pulls either a single, or multiple, messageBundles together to use in a call.  If only one language is required, it would look like this.

def mbl = messageBundleLoader(["en":mbUs])

For multiple languages, or multiple messageBundles, they would be concatenated together with commas.  View a valid list of locales to make sure the parameter in parenthesis is correctly linked to the correct locale.

def mbl = messageBundleLoader(["en":mbUs", "es":mbSpanish])

Throw an Exception (Interrupt Form Save)

Here is where the cool stuff happens.  see post about looping through cells

If a validation error exists, an exception can be generated to stop the form from saving.  To do this, simply use the throwVetoException method.  This accepts 2 parameters.  The first is the messageBundlerLoader, and the second is the id associated to the to be displayed.  Using the example above, and assuming the local is US, the following would stop the form from saving and display a message of  “Only alphanumeric characters can be entered (a-z, 1-9).”

throwVetoException(mbl, "validation.InvalidCharacters")

Consolidated Example

The following example creates two error messages in two languages.  On form save, this will loop through all the cells and throw an error if any value is negative.

def mbUs = messageBundle( ["validation.InvalidCharacters":"Only alphanumeric characters can be entered (a-z, 1-9).",
"validation.Negative":"A positive number is required."])

def mbSpanish = messageBundle( ["validation.InvalidCharacters":"Sólo se pueden introducir caracteres alfanuméricos (a-z, 1-9).",
"validation.Negative":"Se requiere un número positivo."])

def mbl = messageBundleLoader(["en" : mbUs,"es" : mbSpanish])

operation.grid.dataCellIterator.each {  
  if(it.data < 0)  
    throwVetoException(mbl, "validation.Negative")
  }

Wrap Up

It has been a long time since developers have had this kind of control.  The possibilities are only limited by your imagination and business requirements, but there isn’t any validation that can’t be done.  Future posts will tackle validating Run Time Prompts, and taking form validation one step further by adding cell level tool-tips and color coding.

The last thing with these validation calculations is the importance of when they are executed.  The documentation I have from Oracle states something slightly different, so I don’t know if this is the way it is supposed to work, but in my experience, where the rule runs is critical.  Here is what I am experiencing.

  • When the rule is set to Run Before Save, and there is a validation error, the user can’t save the form and an error messages is displayed in the correct locale.  To me, this is the experience that is expected.
  • When the rule is set to Run After Save (which is the way it is documented), and there is a validation error, the user receives an error, but the data is saved.

The difference in the above does provide some interesting options.  Let’s say that we have a form and users are required to allocate an expense.  If the expense is not allocated at 100%, the form can’t be saved.  Assume that there is a rule that the expense shouldn’t be allocated to more than 3 places, but users should be warned if it is.  In this case, if the rule is set to run AFTER save, the user gets the message, but the data is saved.

Either way, if the rule is executed before other rules on the form, no subsequent form will fire if there is a validation error.

 




Adventures in Groovy – Part 12: Learning and Testing Groovy Outside of PBCS

Introduction

For people that are new to Groovy/Java, testing functions that Groovy provides can be a tedious and time consuming process.  Learning anything is.  Trying to do this with the wrong tools compounds it.  I have seen some people give up and walk away from trying to improve applications because they struggle with the Groovy Calculations and the complexity it introduces to go beyond some of the basics, just because they are using a hammer when they need a screwdriver.  For example, it is simple to use a documented example and loop through the cells on a form, but to utilize the Groovy/Java objects and methods is the difference between using the default logic and taking Planning to a whole new level.  For those of us who are learning, testing simple functions can be very painful inside a Groovy Calculation.

I will by preface saying I am not a Groovy developer.  I am learning as I need functionality and I am trying to build a foundation to be as productive as possible.  Although Groovy in PBCS doesn’t give developers full access to all the Java libraries, much of the logic that is needed to develop new functionality can be tested outside of PBCS.  I have found that as I learn more and require more non PBCS related functionality, it is easier to test in the Groovy Console rather than in a PBCS calculation.  Some examples are

  • string functions like replace, regex, concatenate
  • mathematical functions
  • other manipulation that require the use of collections and hash tables

These can be used in looping through grid cells or building evaluation rules on data entered.  Hopefully, this is helpful to those learning Groovy.

How To Get Started

Download Java SDK

Before Groovy can be used, Java has to be installed.  Most systems already have it.  If not, the Java Development Kit can be downloaded and installed.  There is information about which version of Groovy and Java are compatible at groovy-lang.org.  The Java SDK can be downloaded from Oracle.

Download An Editor

Groovy can be edited in many free and paid programs.  Some of them are more robust than others and provide things like automatic code completion, color coding, and more advanced features that aren’t likely required at this novice level.  They also increase the complexity for those that are completely new to writing JAVA or Groovy.  If you are interested in this or need a longer term solution, check out these editors.

For those who want to just get started with a simple and supported editor to test some basic code, try the Apache Groovy Console.  The Windows Installer, the documentation, and the SDK are available to download and install.  Once installed, you are ready to go!  If you go to your Start menu in Windows, you will see a folder for the version of Groovy installed.  In that folder click on Start GroovyConsole to open the editor.

Using The Groovy Console

Much of what is done in the Groovy calculations can’t be accessed here.  We don’t have grids, cells, or any of the PBCS  methods that we interact with in a Groovy Calculation.  Groovy can also access the REST API (outside of Groovy Calculations), which opens up the ability to manage PBCS like EPM Automate.  I recently looped through the product catalogue at BestBuy.com and built a hierarchy!  This is a whole other beast, but it is worth mentioning.

Before we jump into testing a script, here are a few things that will be helpful using the Groovy Console.

  • The editor has two panes.  The top pane is where the script is developed and edited.  The bottom pane is where the results of the script are displayed when it is executed.
  • The toolbar has some common functions.  You can open and save your scripts, redo/undo, and execute from icons in this area.

Examples

I find it very helpful as I am learning, to test the logic and the results in this console.  Once validated, it will be moved to the PBCS calculation and used appropriately.  Here are some examples where it might be useful, and hopefully the separation of where to test what is highlighted.

Regex Example

There was a requirement on a form at a recent client where they wanted to accept input.  They used this to setup properties in the HR system.  The HR system could not accept some characters, so the ask was to only allow alphanumeric characters, a space, an underscore, and a dash.  We had to add validation to the run time prompt, as well as when the data was updated in a form.  Not being an expert with regex, I didn’t want to test this in a calculation (update calc, run calc, open job console, expand status, toggle between windows, etc).

So, I opened the Groovy Console and tested there.  The end result is below, but it was much easier to tweak the regex syntax directly in the console, running it, and seeing the result immediately, in one step.  This was easy to see and verify the output was void of any characters that were not allowed.  The length could be compared, pre and post character removal, and was used to stop the save of the data.

String text = "This - text ! has \\ /allot # of % special % characters"
println text
println text.length()
println text.replaceAll("[^a-zA-Z0-9 _-]", "")
println text.replaceAll("[^a-zA-Z0-9 _-]", "").toString().length()
println text.length() == text.replaceAll("[^a-zA-Z0-9 _-]", "").toString().length()

At this point, I proved out the regex functionality.  I can now go back to the Groovy Calculation and use this logic on the variable returned from the PBCS function (whether it be an RTP or a cell value) and remove the invalid characters or test to see if there are any, and act accordingly.  This is what it would look like

String enteredValue = rtps.RTP_NewEmployee.getEnteredValue();
if(enteredValueAdj.length() == enteredValue.length())
{ 
def mbUs = messageBundle(["validation.InvalidChars":"You have entered invalid characters.  Only alphanumeric characters, spaces, dashes, and underscores are accepted."])
def mbl = messageBundleLoader(["en" : mbUs])
throwVetoException(mbl, "validation.InvalidChars", rtps.RTP_NewEmployee)
}
Converting Nested Collections

I was building a Data Map override from a POV, and it wasn’t validating because some of the variables were collections that included a nested collection.  This whole concept was completely new to me, and again, I didn’t want to have to go through 3-5 steps to see if the result was returning a delimited list of members that the Data Map would accept.  Since I had no initial idea how to accomplish this, I searched and found examples that might accomplish what I wanted to achieve.  It took 5 to 10 iterations of examples to get to what I wanted and understand how this worked.  Updating a script in the Groovy Console, running it, and seeing the results in the same window proved much quicker to find a solution.

In the solution below, I created a variable that replicated the variable that PBCS that was returning (a list).  I was able to build out a few lines to eliminate the nested collections and ported this over to my Groovy Calculation.

This proved out that the simple loop below would give me a list I could pass to the Data Map, and was much quicker to solve than trying to do this in PBCS.

def orig_list = [10, 20, [1, 2, [25, 50]], ['Groovy']]
def usable_list = []

orig_list.collectNested([]) { 
 usable_list << it
 } 
println usable_list 
println '"' + usable_list.join('","') + '"'

The result of the executed script created two lines.  At this point, I could use this function in the Groovy Calculation by replacing the orig_list with the object returned from the PBCS function.  I used the usable_list in the Data Map.

[10, 20, 1, 2, 25, 50, Groovy]
“10”,”20″,”1″,”2″,”25″,”50″,”Groovy”

Wrapping Up

These examples are great examples of how we can use a pair of tools to create business logic efficiently.  If you are a seasoned java developer, much of this might seem ridiculous to you and question why one would ever use something outside of PBCS.  I get it.  Now that I know how these two function work, I likely will not use the Groovy Console to write and test this.  But, as I continue to learn more and more, being able to do this in something outside of PBCS has proven invaluable, increased my productivity, and significantly reduced my frustration.

If you are learning, or are an experienced Groovy developer, please share your insights with the community and post a comment!




Adventures In Groovy – Part 11: Accessing Metadata Properties

Introduction

Groovy opens up a lot of things above and beyond performance improvements and improving the user experience.  One example is the possibility to interact with the metadata.  Dimensions and members can be queried for all types of things which can be useful in many situations.  Is the POV at a level 0?  What is the parent of the current POV member?  Does the member exist in another application?  What about pushing data for specific UDAs and dynamically generating the Data Map?  How about dynamically generating the Data Map to ignore dynamic calculated members?  These are just some examples to get you thinking about where this could be useful.

Code Example

This article won’t get into the logic to accomplish the above examples once the property is identified but will explain how to extract properties for its use.  Below is an example of retrieving every property of an account named Regular_Cases.  This iterates through every metadata property and writes it to the log.

// Get the dimension of the member in question
Dimension AccountDim = operation.application.getDimension("Account")
// Get the member 
Member AccountMbr = AccountDim.getMember("Regular_Cases")
// Print the map to the log
println AccountMbr.toMap()
def memberProps = AccountMbr.toMap()
// Print the member name
println AccountMbr.toString()
// Print every property and corresponding property value
for ( e in memberProps ) {
  println "${e.key} = ${e.value}"
}

When this is executed, the following is sent to the log.

println AccountMbr.toMap() produces

{Formula (rGP)=<none>, Plan Type (GP)=true, Solve Order (rGP)=0, Formula (Fin)=<none>, Data Storage (OEP_WFSC)=never share, Time Balance=flow, Formula=<none>, UDA=HSP_NOLINK, Skip Value=none, Variance Reporting=non-expense, Data Storage (GP)=never share, Essbase Name=Regular_Cases, UUID=c842d186-6d83-4b90-8d1e-49474a6a8a1d, Member=Regular_Cases, Data Storage=never share, Data Storage (rFin)=never share, Formula (rFin)=<none>, Aggregation (rWFP)=+, Formula (GP)=<none>, Data Storage (rWFP)=never share, Data Storage (OEP_REP)=never share, Data Storage (rGP)=never share, Data Type=currency, Formula (OEP_WFP)=<none>, Plan Type (rFin)=true, Aggregation (OEP_WFP)=+, Data Storage (OEP_WFP)=never share, Parent=GP_Accts, Two Pass Calculation=false, Aggregation (GP)=+, Plan Type (rGP)=true, Process Management Enabled=true, Plan Type (rWFP)=false, Source Plan Type=GP, Aggregation (OEP_WFSC)=+, Exchange Rate Type=none, Plan Type (Fin)=true, Alias: English=Regular Cases, Plan Type (OEP_WFP)=false, Aggregation (OEP_REP)=+, Solve Order (rWFP)=0, Data Storage (Fin)=never share, Hierarchy Type=dynamic, Allow Upper Level Entity Input=false, Account Type=revenue, Formula (OEP_REP)=<none>, Aggregation (Fin)=+, Aggregation (rGP)=+, Plan Type (OEP_WFSC)=false, Formula (rWFP)=<none>, Formula Description=<none>, Aggregation (rFin)=+, Solve Order (rFin)=0, Formula (OEP_WFSC)=<none>, Solve Order (OEP_REP)=0, Valid For Consolidations=false, Plan Type (OEP_REP)=false}

for ( e in memberProps ) {println “${e.key} = ${e.value}”} produces

Regular_Cases
Formula (rGP) = <none>
Plan Type (GP) = true
Solve Order (rGP) = 0
Formula (Fin) = <none>
Data Storage (OEP_WFSC) = never share
Time Balance = flow
Formula = <none>
UDA = HSP_NOLINK
Skip Value = none
Variance Reporting = non-expense
Data Storage (GP) = never share
Essbase Name = Regular_Cases
UUID = c842d186-6d83-4b90-8d1e-49474a6a8a1d
Member = Regular_Cases
Data Storage = never share
Data Storage (rFin) = never share
Formula (rFin) = <none>
Aggregation (rWFP) = +
Formula (GP) = <none>
Data Storage (rWFP) = never share
Data Storage (OEP_REP) = never share
Data Storage (rGP) = never share
Data Type = currency
Formula (OEP_WFP) = <none>
Plan Type (rFin) = true
Aggregation (OEP_WFP) = +
Data Storage (OEP_WFP) = never share
Parent = GP_Accts
Two Pass Calculation = false
Aggregation (GP) = +
Plan Type (rGP) = true
Process Management Enabled = true
Plan Type (rWFP) = false
Source Plan Type = GP
Aggregation (OEP_WFSC) = +
Exchange Rate Type = none
Plan Type (Fin) = true
Alias: English = Regular Cases
Plan Type (OEP_WFP) = false
Aggregation (OEP_REP) = +
Solve Order (rWFP) = 0
Data Storage (Fin) = never share
Hierarchy Type = dynamic
Allow Upper Level Entity Input = false
Account Type = revenue
Formula (OEP_REP) = <none>
Aggregation (Fin) = +
Aggregation (rGP) = +
Plan Type (OEP_WFSC) = false
Formula (rWFP) = <none>
Formula Description = <none>
Aggregation (rFin) = +
Solve Order (rFin) = 0
Formula (OEP_WFSC) = <none>
Solve Order (OEP_REP) = 0
Valid For Consolidations = false
Plan Type (OEP_REP) = false
Data Storage (GP) = never share

Getting A Specific Property

Typically, there would not be a need to pull every property.  There might be times when having access to these, however, is useful in calculations.  If a currency calculation is being executed, for example, the rate applied is different if the member is a balance sheet account.  Getting one value can be retrieved by building on the above script.

def keyProp = "Account Type"
if(memberProps[keyProp] = "Revenue"
  {do something}
elseif(memberProps[keyProp] = "Balance Sheet"
  {do something}

Wrap Up

This may seem a little worthless at first, but if you think about all the BSO functions (getting UDAs, Account types for VAR functions, and member relation functions) that require this information, mimicking them in Groovy requires access to the metadata properties.  So, don’t underestimate its use for things like variance, currency, and other calculations, that are done outside of Essbase/Planning calculations and member formulas.