PlanIQ errors and warnings

Some field names are reserved by Amazon Forecast and aren't allowed. Go here for the full list of reserved words.

Change the field names to non-reserved terms.

One of the data columns contains a data type (such as number) that was changed to a different data type.

Revert to the original data type.

Historical/Related data contains an “Exclude value” column for {column name} which doesn’t exist in the module.

Examine your source module and export action, and try again.

All lines in {dataset} export action result file don't contain comma separator. Please check Anaplan Export Action export file format

{Dataset} export action result file format isn't CSV. file type is PDF.

{Dataset} export action result file format isn't CSV.

The prediction target column in the historical data that the forecast uses to base predictions on contains non-numeric values, such as strings or dates.

Review the historical data module and export action for invalid values.

The historical data must contain three columns: item unique identifier, time dimension, and prediction target. The error lists the missing columns.

Missing columns cause issues with the export action.

Review the historical data module and export action to ensure that all three columns are available.

The maximum number of records in historical data is 1,000,000 records.

Reduce the number of records that the export action generates. If possible, use a view with a filter to get a higher level of control.

The historical data module can contain up to 100,000 unique item identifiers. 

Reduce the dataset and train fewer unique items per forecast model.

The historical data contains more than three columns. 

Required columns are: item unique identifier, Time dimension, and prediction target.

Review the historical data module and export action to ensure that only those columns are exported.

More than 50% of the values in the historical data are missing. 

Consider filling in the missing values, or forecast at a different level of granularity.

The prediction target column in the historical data module has no data. 

Review the historical data module and export action to ensure that the data is available.

The related data must have a minimum of two columns: Item unique identifier, and the Time dimension. If any columns are missing, it could be one of these reasons:

• The export action configuration has an issue. 
• Previously available columns in the related data export action are missing. 

Review the related data module and export action.

The related data contains the prediction value column. This can indicate that there's an issue with the export action. For instance, you may have used your historical data export action for the related data. 

Review the export action for the related data.

The related data includes a date format that PlanIQ can't identify. 

Review the date format to ensure that it's compatible with PlanIQ.

The related data contains more than 28 columns (not including the item unique identifier and time dimension). 

Reduce the number of columns in the related data module.

All the numeric and Boolean columns in the related data set are empty. 

Related data must contain an Item unique identifier and Time dimension columns, as well as support columns with numeric or Boolean data.

Review the related data module and export action.

There are no numeric of Boolean columns in the related data set (apart from the item unique identifier and Time dimension column).

Review the related data module and export action.

The attributes data contains more than 10 columns (not including the item unique identifier). 

Reduce the number of columns in the attributes module.

The attributes data doesn't include the item unique identifier column. 

Review that the related data module and export action are configured correctly.

The related data must start on or before the start date of the historical data. This error message is displayed if at least one item’s start date in the historical data is earlier than its first appearance in the related data. 

Adjust the data for the invalid items.

The time scale of related and historical data is incompatible.

You provided invalid forecast frequency: {frequency}, supported: {supported frequencies}.

Historical data can be aggregated, for example from daily to weekly, or from weekly to monthly. Related data can't be aggregated. If you leverage related data the time granularity of forecast models will be aligned to the granularity of the related data.

If related data has a more granular frequency than historic data (such as related data is daily and historic data is weekly), related data can be aggregated in Anaplan modules before being brought into PlanIQ

If you change the granularity of the data, make sure to rerun your data collection.

This can happen if PlanIQ can't detect the time resolution of data. One of the reasons for this could be that the exported dataset is too sparse for PlanIQ to verify the time resolution (daily, weekly, or monthly). 

Omit empty rows for historical data and include empty rows for related data. 

See information on export actions for PlanIQ.

This warning displays if historical data has any of these issues:

• Missing dates
• Duplicate dates
•  dates are not aligned with the time scale of the historical data set. For example, you've specified a weekly scale. All weeks are represented by values for Mondays, and occasionally there's a record with a date on Sunday.

Review the date format in the historical data module and export action.

PlanIQ excluded some columns during the processing of the data collection.

Review the historical data module and export action.

One of the historical data columns contains an integer numeric data type that was later changed to a different data type.

Revert to the original data type.

This warning is shown if the related data has any of these issues:

• Missing dates
• Duplicate dates
• Some dates not aligned with the time scale of the historical data set. For example, you've specified a weekly scale. All weeks are represented by values for Mondays, and occasionally there's a record with a date on Sunday.

Review the date format in the related data module and export action.

At least one item in the historical data set is not in the related data set, or at least one item in the related data set is not in the historical data set.

Review the settings of the export action.

Related data columns listed in the error message are empty. 

Review the related data module and export action.

PlanIQ excluded columns during the processing of the data collection. Exclusion typically happens if you add columns after the data collection was originally created.

Review the related data module and export action.

PlanIQ analyzes the content of columns in the related data and if numeric columns contain non-numeric information, they're ignored.

Review the related data module.

MVLR is the only algorithm that can support more than 13 related columns (in addition to Time dimension and Item unique identifier). 

Review different algorithms to understand the differences between them and update your related data module and export action if needed.

There are items that appear in the historical data that aren't present in the attributes data. 

When you include attribute data as part of the data collection, PlanIQ assumes that items list in the attributes data is the primary list. It ignores all items in historical or related data that are not part of attributes data.

The quantity of items in the data collection you changed is more than 20% of the overall number of items.

PlanIQ won't be able to perform a prediction quality check because the change in the data collection is so dramatic.

Return to your source module and verify the data.

PlanIQ has been unable to check the model prediction quality based on the actuals that have been provided.

The model prediction quality comparison process can't run because a unified actuals timeline can't be constructed.

Forecast action execution and generation of forecasts isn't impacted by this warning.

PlanIQ has been unable to check the model prediction quality based on the previous forecasts.

The model prediction quality comparison process can't run because a unified historic forecast timeline can't be constructed.

Forecast action execution and generation of forecasts isn't affected by this warning.

PlanIQ has been unable to check the model prediction quality.

The model prediction quality comparison process can't run due to an internal error.

Forecast action execution and generation of forecasts isn't affected by this warning.

Your prediction quality has reduced from {previous quality level} to {current quality level}.

The warning message appears when the quality of the forecasting model has decreased compared to its previous performance. 

This overall quality metric is an aggregate metric, which describes the forecast results quality across the entire dataset. If this quality metric warning appears, it indicates the previous quality metric (for example, High) and the current one (for example, Medium), which allows end users to understand the extent of the change.

Forecast action execution and generation of forecasts aren't covered by this warning. The end user should analyze the data or train a new forecast model.