Dos and Don’ts
Topic | Tip | Details |
|---|---|---|
Large data and ModelFormulaContext | ❗ Do not store large data in ModelFormulaContext – this is because it is stored as an XML on the Optimization Model. | |
Do not to run an Optimization logic in the Debug mode | ❗In general, it is not recommended to run an Optimization logic in the Debug mode. It is usually better to create a model, assign the logic to that model and run the logic directly from the model using the Calculate button on each calculation tab. | |
Running a model for the first time | ⚠️ Before running a model with a single-logic configuration for the first time, you need to open the model’s Data Load of the type Model Calculation (it is created automatically when the model is saved) and uncheck the Allow distributed calculation checkbox. Otherwise you will get an error: “Formula for calculation phase not found”. | |
Null values in segmentation dimensions | ❗ Avoid null values in segmentation dimensions. If necessary, create a new field by using an expression which replaces empty values by a text like “Not provided”. | How to Configure Segmentation-Optimization Models (Vesper 6.2 and higher) |
Model evaluation on line item level | ⚠️ Be aware that running model evaluation on the line item level can negatively impact performance. Each model run may add 0.5-1sec for every line item, which results in quote being recalculated for e.g. 30 seconds or more. This may not meet the end users’ expectations when it comes to application performance. |
Tips and Notes
Topic | Tip | Details |
|---|---|---|
Model name length limitation | For a model name, there is a limit of 50 characters. (This limit is valid for all FC uniqueName values in Pricefx and it originates from the limit of 63 characters in the underlying Postgres database.) | How to Configure Segmentation-Optimization Models (Vesper 6.2 and higher) |
Set heartbeat | It is recommended to set your partition's heartbeat to a low value (20 s). |