1

Adventures in Groovy – Part 52: And You Thought Essbase Only Stored Numbers

My 20+ years of using Essbase I was told, and had no reason not to believe, only stored numbers. We obviously have lists and with Planning it appears we have text. If you aren’t familiar with how this works, the strings are stored in the Planning repository and the index is stored in Essbase. If you thought the same thing, you were as wrong as I was.



What is NaN

I have been learning and implementing Groovy solutions for 2-3 years now and came across something I now only have never seen, but didn’t think was possible. Java, and therefore Groovy, has a concept of NaN. NaN stands for Not A Number. NaN is the result of mathematical operators that create non numbers. Log, square root, division, and I am sure plenty of other formulas that I learned before I was 15 and long forgot, can result in what Java interprets as non numeric values. The two that I have found are NaN and Infinity. An example of 4/0 would result in NaN. 0/4 would result in Infinity.

NaN in Groovy

Prior to about 2 months ago, I accounted for these scenarios in my logic and never had an issue. Recently, in writing some basic math, like revenue / units, I didn’t account for the possibility that revenue or units would be a zero. If these scenarios are tested in Groovy, errors are encountered and honestly, I thought my logic in a business rule would have produced a divide by 0 error when the denominator was a 0.

java.lang.ArithmeticException: Division by zero
at ConsoleScript2.run(ConsoleScript2:1)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

I thought, like in Essbase, 4/0 would result in a 0. I found out the hard way that is not the case!

Types Of NaNs and Infinities

In my case, I didn’t care of the sub type of Nan or Infinity the results was, just that it happened and I needed to account for it. These can be checked very simply.

double simpleSample = 4/0
if( simpleSample.inNaN() || simpleSample.isInfinite() ) {
  println 'ERROR'
}

Your situation might be different. If it is, these are the types I am aware of that you can check for

//NaN variances
isNaN(nan)
isNaN(zero_div_zero)
isNaN(sqrt_negative);
isNaN(inf_minus_inf);
isNaN(inf_times_zero);
isNaN(quiet_nan1);
isNaN(quiet_nan2);
isNaN(signaling_nan1);
isNaN(signaling_nan2);
isNaN(nan_minus);
isNaN(log_negative);
isNaN(positive_inf);
isNaN(negative_inf);
 // Infinite variances
isInfinite(positive_inf);
isInfinite(negative_inf);

What Do You See In Planning and Essbase

So here is where I really was confused! Everything I I thought I knew was wrong.

Surprise Number One

If either of these conditions occurs, the cell that was calculated in Groovy and stored in Planning/Essbase is actually stored differently. I can’t say for sure what happens on the back end, but when the data is exported, rather than a numeric value, it will export NaN. Yes, you will see something like 10,20,20,NaN,40….

Surprise Number Two

If either of these conditions occurs, the cell that was calculated in Groovy and stored in Planning/Essbase shows a number that makes no sense in a data form when opened in Smart View. A value of 65535 will be displayed. This value can be edited/changed. If it is the source of another member formula or calculation, it will also show a value of 65535.

Surprise Number Three

The same thing is NOT what you see in a data form opened in the UI. In the UI (web version), NaN or Infinity will actually be displayed in the effected cell. This almost makes sense if I didn’t see 65535 in Smart View.

Stop NaNs From Happening

There are probably a million ways to handle this. For what it is worth, I want to share how I handled it and why. First, I created a function in my calculation that accepted one parameter, which was the value in which I was evaluating for Nan or Infinity. Inside this I used an Elvis operator and returned 0 if it was Nan or Infinity, and the value submitted to the function if it was a numeric value. The reason I created a function was because I had more than 30 formulas that I needed to check for this and it was easier to write the code once.

double nanCheck(double input){ (input.isNaN() || input.isInfinite()) ? 0 : input }

// Use Case Example
DataCell rate
DataCell units
operation.grid.dataCellIterator({DataCell cell -> cell.edited}).each{cell->
    rate = cell
    if(rate.accountName == 'Small_Unit_Cost'){
        units = cell.crossDimCell('Small_Units')
        units.data = nanCheck(cell.crossDimCell('Revenue').data / rate.data)
    }
    else if(...)
        {...}
}

That’s A Wrap

it is really important to account for this for obvious reasons. If you are testing for NaN and Infinity, save yourself some trouble and if there is a possibility of it occurring, start doing it now. It is a pain to strip it out afterwards if it gets into a UAT situation or even Production. One last thing. If you are looking at this and thinking, this should really return #Missing. You surely can do that. There are a few changes that have to be made. First, the function can’t be double. Since #Missing is a string, it would need to be a string. The second issue is that you can’t set data, which is a double, to a string. You would have to use formattedValue. The changes would look something like this.

String nanCheck(double input){ (input.isNaN() || input.isInfinite()) ? '#Missing' : input }

// Use Case Example
DataCell rate
DataCell units
operation.grid.dataCellIterator({DataCell cell -> cell.edited}).each{cell->
    rate = cell
    if(rate.accountName == 'Small_Unit_Cost'){
        units = cell.crossDimCell('Small_Units')
        units.formattedValue = nanCheck(cell.crossDimCell('Revenue').data / rate.data)
    }
    else if(...)
        {...}
}



Upgrade Or Downgrade To Or From Hybrid In The Cloud

There are benefits to moving to Hybrid, but there are also some challenges.  The content of this post is not around the pros and cons, but the fact that you can upgrade your environment to use it.  If you find it isn’t for you, you can “downgrade” back to BSO.  The flexibility provides everybody the ability to try it.



Recreate Introduction

EPMAutomate comes with a function that allows the ability to restore an environment to a clean slate.  I don’t think this is new to anybody that has used EPMAutomate or EPM Cloud Planning.  What might be a surprise is that it does more than just reset an environment so you can start over.  It can also:

  1. Change the type of Essbase database to Hybrid or a standard BSO.
  2. Temporarily convert a Planning, Enterprise Planning, Tax Reporting, or Financial Consolidation and Close environment to an Account Reconciliation, Oracle Enterprise Data Management Cloud, or Profitability and Cost Management environment.

Using Recreate

The usage of the Recreate command is as follows, which all options.

 epmautomate recreate [-f] [removeAll=true|false] [EssbaseChange=Upgrade|Downgrade] [TempServiceType=Service_type]
  • -f forces the re-create process to start without user confirmation. If you do not use the -f option, EPM Automate prompts you to confirm your action. Be careful using this option.  If you have a long day and aren’t focused, this can make the day a whole lot worse!
  • removeAll, removes all of the existing snapshots, as well as the content of the inbox and outbox.  The default is false, meaning it retains the snapshots and the content of inbox and outbox and nothing is removed.
  • EssbaseChange upgrades or downgrades the current Essbase version in legacy Oracle Financial Consolidation and Close Cloud, Oracle Enterprise Planning and Budgeting Cloud or Planning and Budgeting Cloud Plus 1 environments.
  • TempServiceType temporarily converts an environment to a different service environment.

Changing Your Essbase Version

To change your environment to BSO from Hybrid

epmautomate recreate EssbaseChange downgrade

To change your environment to Hybrid from BSO

epmautomate recreate EssbaseChange upgrade

Trying A Different Service

There are some details that must be understood to use this option and is dependent on the version of the cloud service you have.  For subscriptions other than EPM Standard Cloud Service and EPM Enterprise Cloud Service, meaning PBCS and EPBCS, you can use this option to convert, temporarily, to

  • Account Reconciliation
  • Oracle Enterprise Data Management Cloud
  • Profitability and Cost Management environment

To use this option to convert your environment to something it wasn’t originally intended for:

epmautomate recreate -f removeAll=true TempServiceType=ARCS

To change your environment back to its original service:

epmautomate recreate

For EPM Standard Cloud Service and EPM Enterprise Cloud Service subscriptions, you can use this option to convert to any supported EPM Cloud service.  EPM Enterprise Cloud Service subscriptions use a common EPM Cloud platform. Initially, you can deploy any supported EPM Cloud business process. 

To switch from a deployed business process to another, you must re-create the environment to delete the current deployment and to bring it back to the original EPM Cloud platform. You then re-create it again as the new service type.

For example, if you created an Account Reconciliation business process but now want to create an Oracle Enterprise Data Management Cloud environment, you must run the re-create command twice.

First, reset the service.

epmautomate recreate -f removeAll=true

Second, change the service type.

epmautomate recreate -f TempServiceType=EDMCS

The acceptable service types, currenty, are

  • ARCS (Account Reconciliation)
  • EDMCS (Oracle Enterprise Data Management Cloud)
  • EPRCS (Narrative Reporting)
  • PCMCS (Profitability and Cost Management)

That’s A Wrap

It is great that Oracle allows us to do these things. We have a ton of flexibility, not normally afforded to us in the cloud, to test and use different core database types.  It also allows those using the old SKU to try the new services, or business processes, to see if they might be something you want to purchase. 

If you want to give Hybrid a try, use your test environment and give it a spin.  If you want to get exposed to one of the other business processes, you now have the ability to see it without jumping through hoops.




Working With Planning Formula Expressions

Most of us know that there is a button in the calc rule editor that allows us the ability to select a smart list and the smart list entry.  It will add something [[smartlist name.smartlist entry]].  If this is new to you, what it does is replace reference the smart list and replaces it with the numeric value that exists in Essbase.  The beauty of this is that it is dynamic, so if the smart list is changed in any way, you don’t have to go into your rules and replace change the index values for the smart list entries to match.  Guess what, there are more!



Well, don’t I feel like the F*@$& idiot, to pull a quote from A Few Good Men.

What Is A Planning Formula Expression

As described above, it is an expression that allows you to get valuable information dynamically about artifact properties in a Planning application.  The following formula expressions currently exist.

  • SmartLists
  • Dimensions
  • Planning User Variables
  • Periods
  • Scenarios
  • Cross-References
  • Workforce Cube Year to Date
  • Get ID for String

SmartLists

I already discussed the Smart List, but here is an example if this is new to you.  The calculation manager syntax is [[SLName.entryname]].

FIX (Mar, Actual, Working, FY15, P_000, "111") 
   "Product Channel" = [[Channel.Retail]] ;
ENDFIX

Which would return something like this.

FIX (Mar, Actual, Working, FY15, P_000, "111") 
   "Product Channel" = 2 ;
ENDFIX

Dimensions

The dimension expressions are not all that useful unless you are building calculations that might go across applications that have different names for the 6 required dimensions, plus currency.  Using the following dimension tags, the customized name will be returned when they are added to the dimension expression. The syntax used for this function is [[Dimension("DIM_NAME_ENTITY")]].

  • DIM_NAME_PERIOD
  • DIM_NAME_YEAR
  • DIM_NAME_ACCOUNT
  • DIM_NAME_ENTITY
  • DIM_NAME_SCENARIO
  • DIM_NAME_VERSION
  • DIM_NAME_CURRENCY

An example would look like this. This runs a calc dim on whatever your account dimension is.

CALC DIM([[Dimension("DIM_NAME_ENTITY")]]);

In this application, Entity is named Entity, so the above script returns:

CALC DIM ("Entity");

If the entity dimension was named Cost Center, it would return:

CALC DIM ("Cost Center");

Planning User Variables

Planning user variables return the user variable’s member.  This can be pretty useful if you have variables that are used to do things like fix on their area of a hierarchy.  These can be gathered through run time prompts if they exist in the POV, but that isn’t always the case.  You might use them to show the products, for example, that are under a user variable, in the rows, in which case without Groovy, it can’t be passed in RTPs. The calculation manager syntax is [[PlanningFunctions.getUserVarValue("xyz")]].

An example where the user variable is used to run a calculation might look like this.

FIX (Feb, Actual, Working, E_000, @RELATIVE([[PlanningFunctions.getUserVarValue("Product View")]],0) )
   Revenue = Units * Cost;
ENDFIX

Period Functions

Period(periodName) returns the specified period. The options for this function are and the calculation manager syntax is [[Period("FIRST_QTR_PERIOD")]].

  • FIRST_QTR_PERIOD
  • SECOND_QTR_PERIOD
  • THIRD_QTR_PERIOD
  • FOURTH_QTR_PERIOD
  • FIRST_PERIOD
  • LAST_PERIOD

This example:

FIX ( Mar, Actual, Working, P_000, "6100", FY15 )
   "120" =[[Period("FIRST_QTR_PERIOD")]];  
ENDFIX

would return a script like this

FIX (Mar, Actual, Working, P_000, "6100", FY15) 
   "120" = "Mar";
ENDFIX

The NumberofPeriodsInYear returns the number of periods in the year and NumberofYears returns the number of years in the application. The calculation manager syntax for this is

  • [[NumberOfPeriodsInYear]]
  • [[NumberOfYears]]

The following example

FIX (Mar, Actual, Working, P_000, "6100", FY15)
   "120"=[[NumberOfPeriodsInYear]];
   "120"=[[NumberOfYears]]; 
ENDFIX

would produce this.

FIX (Mar, Actual, Working, P_000, "6100", FY15)
   "120"=12;
   "120"=9;
ENDFIX

Scenarios

This one is my favorite ones.  I have been using Groovy to get these not knowing they existed.  These allow the reduction of if statements and improved performance.  We can get the open periods.  If this go across years, then my groovy solution probably comes back into play.  For ranges that include one year, or even two could be handled, this offers great functionality.  The options available are

  • Start Year
  • End Year
  • Start Month
  • End Month

The calculation manager syntax is as follows.

  • [[getStartYear(“ScenarioName”)]]
  • [[getEndYear(“ScenarioName”)]]
  • [[getStartMonth(“ScenarioName”)]]
  • [[getEndMonth(“ScenarioName “)]]

A use case would look something like this. Assume{rtpScenario} is a run-time prompt variable of type member with a default value of “Actual”:

FIX({rtpScenario}, [[getStartYear({rtpScenario})]]:[[getEndYear({rtpScenario})]],
   [[getStartMonth({rtpScenario})]]:[[getEndMonth({rtpScenario})]])
      FIX ( Working, P_000, "111")
        "5800" = 5500;
      ENDFIX
ENDFIX  

This would build out the following calculation

FIX ("Actual", "FY10" : "FY18", "Jan" : "Dec") 
    FIX (Working, P_000, "111") 
      "5800" = 5500;
    ENDFIX 
ENDFIX

If your open range consisted of two years, you could do something like this

FIX({rtpScenario},[[getStartYear({rtpScenario})]],[[getStartMonth({rtpScenario})]]:”Dec”)
      FIX ( Working, P_000, "111")
         "5800" = 5500;
      ENDFIX
ENDFIX  
FIX({rtpScenario},[[getEndYear({rtpScenario})]],”Jan:[[getEndMonth({rtpScenario})]]
      FIX ( Working, P_000, "111")
        "5800" = 5500;
      ENDFIX
ENDFIX  

Cross-References

This function comes in a few flavors but does something pretty awesome. How it works might change your naming convention a little, or make it more consistent anyway. What id does is generate a cross dimensional reference to our default members, like No Product. The syntax is CrossRef(accountName, prefix, true) but the last two parameters are optional. If you use CrossRef(“Revenue”), it would produce the following, assuming your 6 required dimensions and a product dimension.

"BegBalance"->"No Scenario"->"No Version"->"No Entity"->"No Product"->"Revneue";

I can change my prefix by adding the second parameter. I don’t like having spaces in my member names, so I would do the above with CrossRef(“Revenue”,”No_”) which would produce

"BegBalance"->"No_Scenario"->"No_Version"->"No_Entity"->"No_Product"->"Revneue";

If I change my syntax to CrossRef(“Revenue”,”No_”,true) I have a cross dim operator for all dimensions Except Period (uses BegBalance), and Currency, but includes year

"BegBalance"->"No_Year"->"No_Scenario"->"No_Version"->"No_Entity"->"No_Product"->"Revneue"; 

Used in a Fix Statement the following example

FIX (Aug, Actual, Working, FY15, P_000, "112")
   "111" = [[CrossRef("5800", "No_", true)]];
ENDFIX

would produce the following script.

FIX (Aug, Actual, Working, FY15, P_000, "112") 
   "111" = "BegBalance"->"No_Year"->"No_Scenario"->"No_Version"->"No_Entity"->"No_Product"->"5800";
ENDFIX

Workforce Cube Year to Date

If you use workforce, you probably have noticed the members it generates to get the month index for both the calendar and fiscal period. These can be used with this function to build a calendar to date value. The syntax is [[CYTD(memberName)]]. If you have renamed “Cal TP-Index” and “Fiscal TPIndex,”, then you have to specifically name the members in two additional parameters and the syntax is [[CYTD(memberName, calTpIndexName, fiscalTPIndexName)]]. This method is really easy to use and looks like this.

Fix (NOV, Actual, Working, FY15, P_000, "112")
   "5800" = [[CYTD("6100")]];
ENDFIX

If the default names are changed, it would look a little different.

Fix (Dec, Actual, Working, FY15, P_000, "112")
   "5800" = [[CYTD("6100", "Cal TP-Index", "Fiscal TPIndex")]];
ENDFIX

Get ID for String

This doesn’t solve all the problems around Smart Lists and text accounts, but it is a step in the right direction. If you don’t know, both of these are held in the Planning repository. Essbase ONLY stores numbers. Look at this as the index to the value you see in Planning. The repository has the map from index to value. In an Essbase calculation, you can’t set a text account to a text value. Well, actually, you can. The syntax for this function, which assigns a text value, is [[PlanningFunctions.getIdForString("text")]]. This allows you to set the value of a text account to a string.

In Planning, you have an account named “acct1 text” that is of type text. You want to copy your values from FY16 Dec to FY17 Mar, and change the text account to “Not Budgeted,” it would look like this.

FIX (Actual, Working, P_000, "210")
   DATACOPY FY16->Dec TO FY17->Mar;
   Mar(
      "acct1 text"->FY17 = [[PlanningFunctions.getIdForString("Not Budgeted")]];
   )
ENDFIX

That’s A Wrap

One last thing. If you use any of these in a member formula, for some reason you have to remove a bracket on each side. So, instead of two, you just need one. I will say I have not tested all of these, but the ones I have tested/used do follow this pattern. Hopefully Oracle keeps expanding these. Although they aren’t as helpful as they were prior to Groovy, they are simpler to use than implementing a Groovy solutions for some of these needs. For you lifers, it is things like this that a newb tells you. Don’t ever think you can’t learn from somebody that “knows nothing.”




Adventures in Groovy – Part 51: Dynamically Changing Storage Properties When Using Hybrid

With hybrid being used more and more there is a need to manage the storage methods of different levels of sparse dimensions.  Whether it is a staggered hierarchy or not, getting the storage method from the source can sometimes be challenging.  More often times than not, you may want to own it on the Planning side so you can change it at will and not have to go through the typical IT change order process that may take weeks, or even months, to go through the full development cycle.



Managing this manually would not be fun, especially if the hierarchy is loaded more often than monthly.  Yes, you could use the Smart View admin option, but it is manual and let’s face it, you have a ton going on and you will make mistakes. 

In Comes Groovy

With Groovy, a calculation can be written to update metadata.  I have talked about this in several other posts, but I am going to walk through a couple specific examples that are for specific situations.  I think this will spark some interest of taking this further for situation similar, or completely different. 

Reusable Concepts

Before I jump into the situations and examples, there are a couple techniques that will be reused in all the examples.  Rather than repeatedly explain them, let me first introduce them.

First, this situation assumes that the storage methods are different for the plan types.  This might be more unique, but it is easy to deal with.  If this isn’t the case, the properties in the example can be change to “Data Storage”

It is always a good idea to start every Groovy script off with the RTPS tag.  To understand more about why this is important, read Part 49  This will be used in each example.

/*RTPS: */

Each example requires methods that have to have the dababase passed to it.  The easiest way to get the cube the rule runs on is to use rule.cube.  There are other ways to accomplish it, but this is the shortest and most dynamic.

List<Member> products = operation.application.getDimension("Product",rule.cube).getEvaluatedMembers("Descendants(Product)", rule.cube)

Each example gets the dimension and holds it in a variable.  The method requires a pointer to a cube, or cubes.  Often it is easier to pass the cubes in the application, rather than one cube, to make sure all artifacts are available and not hard coded.  operation.application.cubes as Cube[] returns all the plan types as an array of variables that are of type cube.

A note about the parameters that can be used.  It is much faster to use the same parameters that are used in planning, like the options in a data map.  You CAN use most of the Essbase function.  Oracle doesn’t recommend them.  They are slower, but if you are not iterating and running the request numerous times, I haven’t noticed a difference.  In this example, it is executed once, so the performance degradation is minimal.

Dimension objDim = operation.application.getDimension('Product',operation.application.cubes as Cube[])

To get and set the properties of a member, the toMap method is used.  This will return all the properties of the member and I wrote a summary of the use of this method in a prior post found in this post – Part 11 – Accessing Metadata

Map<String,Object> memberProps = it.toMap()

Lastly, if you aren’t familiar with regular expressions, they can be of great use.  I have a module dedicated to this in xxxxx.  I struggled understanding regular expressions for years.  But I promise you, if you take 4 hours and focus on learning them, it will click.  To use it in Groovy, using the matches method allows this.  Briefly, here are some basic concepts.  A ^ means starts with.  A $ means ends with. A dot means any character, and following that with an asterisk means many.  So .* means one to many characters of any type

.matches("^.*Region$") || it.name.matches("^District.*$")

Setting All Parents To Dynamic

If you have a smaller hierarchy, one with maybe only a few levels, it might be advantageous to just set all the parents to dynamic.  The following script iterates through all the product members and sets every parent to dynamic.

/*RTPS: */
List products = operation.application.getDimension("Product",rule.cube).getEvaluatedMembers("Descendants(Product)", rule.cube)
List<Member> lev0Products = operation.application.getDimension("Product",rule.cube).getEvaluatedMembers("ILvl0Descendants(Product)", rule.cube)
Dimension objDim = operation.application.getDimension('Product',operation.application.cubes as Cube[])
products.each{
try{
       Map<String,Object> memberProps = it.toMap()
       if(lev0Products.contains(it)){
           memberProps["Data Storage (${rule.cube})".toString()] = 'never share'
    }
    else{
           memberProps["Data Storage (${rule.cube})".toString()] = 'dynamic calc'
    }
    objDim.saveMember(memberProps)
}
catch(Exception e) {
    println("Exception: ${e}")
    println it.name
}
}

Use Patterns To Set Parent Storage Property

In some situations, there are patterns to the levels of your hierarchy.  Maybe you have regional levels that are definable and unique that can be used to set different levels to dynamic.  Assume the following naming convention for this example

  • Total Products
    • West Region (everything ends in Region)
      • District 1 (everything starts with District)
/*RTPS: */
// Get every product in the hierarchy
List<Member> products = operation.application.getDimension("Product",rule.cube).getEvaluatedMembers("IDescendants(Product)" , rule.cube)
// Assign the producdt dimension to a variable
Dimension objDim =  operation.application.getDimension('Product',operation.application.cubes as Cube[])
// Loop through each product
products.each{
       // if the product matches these expressions, change the app setting to dynamic
       if(it.name.matches("^.*Region$") || it.name.matches("^District.*$") || it.name == 'Total Products'){
              Map<String,Object> memberProps = it.toMap()
        memberProps["Data Storage (${rule.cube})".toString()] = 'dynamic calc'
        objDim.saveMember(memberProps)
       }
       // otherwise change it to never share
    else{
              Map<String,Object> memberProps = it.toMap()
        memberProps["Data Storage (${rule.cube})".toString()] = 'never share'
        objDim.saveMember(memberProps)
    }
}

More Complicated Possibilities

There are a bunch of other possible needs.  Let’s say you have a need to make everything above level 3 dynamic.  First, if the hierarchy is staggered, the same level can be a 1 and 5.  You would have to decide how to handle that.  I would lean toward if it was a level 1 and a 5, I would make it dynamic because that might also mean your level 5 and 9 in that portion of the hierarchy would be a pretty deep hierarchy to make 9 levels dynamic.  Every situation is different, and performance would have to be evaluated, but the complexity of identifying how to set the storage in these situations is what I am trying to explain.

If you want to use patterns, you may also want to ensure that the pattern isn’t replicated at a parent and level 0, so there may be a need to check for both a pattern and the level of the member.

Obviously, there are an infinite amount of possibilities and each one could introduce complexity.  Just understand that almost anything can be defined by patterns and levels and can be accomplished, but the level of complexity of your logic or your regular expression may increase.

That’s A Wrap

The bottom line is that we now have the ability to do a lot of things we relied on the source system to do.  Or, maybe external scripts were run using Perl, or VBScript, or PowerShell.  We can use metadata properties, dynamic levels, any other repeatable pattern that might possibly come up.  It is fast and reliable, and completely in our control so we don’t rely on others when it is broken or needs changed.