Introduction
SAP planning developers can implement planning logic in SAP Analytics Cloud (SAC) by building data actions or multi-actions. Unfortunately, there are also situations where these tools are too limited. With the introduction of API steps and data loads in the multi-actions we can go beyond SAC to make our data calculations more powerful and establish more comprehensive planning flows. The blog series “Extending SAC planning” beautifully describes a practical example of how to set this up.
In this blog, I will investigate if it is possible to determine an approach for a SAC planning project to establish our planning logic for a few different use cases. Our first question always should be whether the SAC planning logic is adequate for our requirements, and if not, whether we should go for backend HANA modeling capabilities. This would allow us to use more advanced calculations like calculated columns, joins, and SQL window tables. And what about a hybrid way, where we combine data actions & multi-actions with HANA modeling capabilities?
I will investigate two use cases:
Analysis
I will analyze the use cases from three perspectives.
In the standalone use case, the native HANA database executes all our calculations made in SAC. In that way it ensures seamless execution of calculations without the replication of data. No additional systems are required and the user should only have a SAC planning license to execute multi actions & data actions.
In the hybrid case, we encounter a more complex setup as we need to utilize a DP agent to access tables from SAC and make them visible in SAP Datasphere. That introduces potential calculation delays and a more complex system setup by establishing connections between SAC, the DP-agent, and Datasphere. Furthermore, a user must have an additional license to perform Datasphere calculations.
Let's explore the various types of data logic that we can use in standalone and hybrid cases. The image also demonstrates how we can utilize REST API within SAC multi-actions, although this will be out of the scope of this analysis.
The image shows what types of calculations we can make in the data actions & multi-actions. SAC provides standardized calculations such as currency conversion, allocation & predictive steps. It makes sense to use these standardized objects as it requires more effort to build in Datasphere. However, it's important to note that these standardized logics may be limiting when it comes to more complex use cases.
For these use cases, we rely on advanced formulas step in the data actions that offer more flexibility. This scripting environment allows us to utilize logic like:
The downside is that this scripting language is specific to SAC and not used in different IT tools. Less knowledge sharing and resources are available online compared to other scripting languages. Considering these factors, it raises the question of whether similar tasks can be accomplished in Datasphere using SQL-based views. The SQL language has more advanced functionalities, a largere knowledge base, and provides a visual user interface. We answer this question after discussing the optimization flows in the next paragraph.
In situations where advanced formulas steps are not applicable to meet the requirements, we definitely need more powerful calculations to accomplish our tasks. To mention a few use cases where the hybrid planning logic comes into play.
By adopting a hybrid planning approach, we can address these limitations and achieve more robust and flexible calculations by leveraging the power of SQL views and integrating them with our SAC planning data.
Next, let's explore the optimizations we can achieve with our calculations to improve execution speed and minimize system impact. I'll examine the optimizations within SAC itself and then consider how they apply to the hybrid use case.
We begin with a physical SAC model with data where the recommended planning area (RPA) is enabled for optimization.
In this way, several optimizations are implemented by calculating numbers within SAC.
In the hybrid case, we use the same RPA and DAC settings as in the standalone case.
In my opinion, this is a significant drawback of this approach, as it impacts the performance of the SAC system and increases waiting times for end-users to see the complete results.
Conclusion
In conclusion, when considering our standalone use case with SAC, it appears to be the preferred option for executing planning logic as it provides various SAC optimizations.
I previously questioned whether the hybrid logic could serve as an alternative to advanced formulas steps, given their increased power and familiarity with SQL. However, the hybrid solution cannot leverage optimizations such as delta loads, leading to increased system load and requiring more processing power. As a result, it is likely not a favorable alternative to the optimized advanced formulas steps.
Nevertheless, in complex scenarios where SAC as a standalone tool falls short, the hybrid proposition can be intriguing as it enables the implementation of more sophisticated planning logic if necessary, albeit at the expense of optimization. It is important to recognize the trade-off and accept a less streamlined process when opting for the hybrid approach.
SAP planning developers can implement planning logic in SAP Analytics Cloud (SAC) by building data actions or multi-actions. Unfortunately, there are also situations where these tools are too limited. With the introduction of API steps and data loads in the multi-actions we can go beyond SAC to make our data calculations more powerful and establish more comprehensive planning flows. The blog series “Extending SAC planning” beautifully describes a practical example of how to set this up.
In this blog, I will investigate if it is possible to determine an approach for a SAC planning project to establish our planning logic for a few different use cases. Our first question always should be whether the SAC planning logic is adequate for our requirements, and if not, whether we should go for backend HANA modeling capabilities. This would allow us to use more advanced calculations like calculated columns, joins, and SQL window tables. And what about a hybrid way, where we combine data actions & multi-actions with HANA modeling capabilities?
I will investigate two use cases:
- The standalone use case:SAC tenant with data actions & multi actions planning logic.
- The hybrid use case: SAC tenant with data actions & multi-actions combined with Datasphere remote SQL views.
Analysis
I will analyze the use cases from three perspectives.
- The complexity of setup: the prerequisites needed and the need for data replication or not;
- Power of calculations: Ability to create more comprehensive planning flows;
- Calculation optimizations: Ability to use optimization to deliver faster results to end users.
The complexity of setup
| Standalone | SAC | Hybrid | SAC & Datasphere | |
| Place of data logic execution | Data logic in SAC HANA database | Data logic via Remote tables in Datasphere |
| Prerequisites | SAC instance | SAP Datasphere instance |
| Planning license needed | Datasphere license needed | |
| Data provisioning agent setup |
In the standalone use case, the native HANA database executes all our calculations made in SAC. In that way it ensures seamless execution of calculations without the replication of data. No additional systems are required and the user should only have a SAC planning license to execute multi actions & data actions.
In the hybrid case, we encounter a more complex setup as we need to utilize a DP agent to access tables from SAC and make them visible in SAP Datasphere. That introduces potential calculation delays and a more complex system setup by establishing connections between SAC, the DP-agent, and Datasphere. Furthermore, a user must have an additional license to perform Datasphere calculations.
2 Power of calculations
Let's explore the various types of data logic that we can use in standalone and hybrid cases. The image also demonstrates how we can utilize REST API within SAC multi-actions, although this will be out of the scope of this analysis.
SAC-based planning logic
The image shows what types of calculations we can make in the data actions & multi-actions. SAC provides standardized calculations such as currency conversion, allocation & predictive steps. It makes sense to use these standardized objects as it requires more effort to build in Datasphere. However, it's important to note that these standardized logics may be limiting when it comes to more complex use cases.
For these use cases, we rely on advanced formulas step in the data actions that offer more flexibility. This scripting environment allows us to utilize logic like:
- If/Then
- For loops
- Variables
- Value lookups
- Delete, append, and overwrite database records
The downside is that this scripting language is specific to SAC and not used in different IT tools. Less knowledge sharing and resources are available online compared to other scripting languages. Considering these factors, it raises the question of whether similar tasks can be accomplished in Datasphere using SQL-based views. The SQL language has more advanced functionalities, a largere knowledge base, and provides a visual user interface. We answer this question after discussing the optimization flows in the next paragraph.
Hybrid planning logic
In situations where advanced formulas steps are not applicable to meet the requirements, we definitely need more powerful calculations to accomplish our tasks. To mention a few use cases where the hybrid planning logic comes into play.
- If we need to link information from multiple models, the "Link models" functionality of advanced formulas steps may prove to be restrictive. We cannot use different joins between models, which is possible with SQL graphical views.
- Advanced formula steps have limitations when calculating based on dimensional properties. It becomes challenging to group multiple members with the same properties and then use different properties from another dimension within the same calculation.
- When making calculations based on real-time data, advanced formulas steps fall short as they rely on data stored in physical tables within SAC and not support live models. However, with a hybrid use case, we can utilize remote tables to perform calculations on the live models.
By adopting a hybrid planning approach, we can address these limitations and achieve more robust and flexible calculations by leveraging the power of SQL views and integrating them with our SAC planning data.
3 Calculation optimizations
Next, let's explore the optimizations we can achieve with our calculations to improve execution speed and minimize system impact. I'll examine the optimizations within SAC itself and then consider how they apply to the hybrid use case.
SAC-based planning logic
We begin with a physical SAC model with data where the recommended planning area (RPA) is enabled for optimization.
- Based on the user's writing rights specified by Data Access Controls (DAC), SAC creates a writing table when the user edits any data.
- This writing table is the basis for data action calculations, allowing them to be executed on a specific selection rather than the entire model's data.
- Any changes made by the user are recorded in a delta table, which speeds up the publishing process by limiting the action to only the delta records.
In this way, several optimizations are implemented by calculating numbers within SAC.
Hybrid planning logic
In the hybrid case, we use the same RPA and DAC settings as in the standalone case.
- Before the changes are visible in the remote table within Datasphere, we need to publish the user's input data first.
- Then, the DP agent creates a remote table in Datasphere, where calculations are performed using a SQL view. This introduces a delay compared to the standalone option.
- Afterward, the calculated data needs to be imported via the data load mechanism into SAC. The whole remote table needs to be initialized by SAC without the possibility of using delta loads based on updated data leading to further delays in the importing process.
In my opinion, this is a significant drawback of this approach, as it impacts the performance of the SAC system and increases waiting times for end-users to see the complete results.
Conclusion
In conclusion, when considering our standalone use case with SAC, it appears to be the preferred option for executing planning logic as it provides various SAC optimizations.
I previously questioned whether the hybrid logic could serve as an alternative to advanced formulas steps, given their increased power and familiarity with SQL. However, the hybrid solution cannot leverage optimizations such as delta loads, leading to increased system load and requiring more processing power. As a result, it is likely not a favorable alternative to the optimized advanced formulas steps.
Nevertheless, in complex scenarios where SAC as a standalone tool falls short, the hybrid proposition can be intriguing as it enables the implementation of more sophisticated planning logic if necessary, albeit at the expense of optimization. It is important to recognize the trade-off and accept a less streamlined process when opting for the hybrid approach.
- SAP Managed Tags:
9 Comments
