Step 1 - Create the Anaplan Audit Datasets

🔍 What You’re Doing

With your Anaplan connection in place, you now create the Nexadata datasets that will pull the raw audit data from Anaplan. Each dataset maps to a different Anaplan Export Type, giving you the Events feed plus the lookup datasets (Users, Workspaces, CloudWorks Jobs, Models, Imports, Exports, Processes, and Actions) needed to resolve IDs into human-readable names.

These datasets are the foundation the rest of the guide builds on. The workflow template you import in Step 3 configures Pipelines only, so it creates none of these for you. Every dataset it expects has to exist first, with the right columns already in place, which is why this comes before the import.

Note

Make sure you’ve completed the prerequisites listed in the Overview before starting this step.

---

✅ Summary of Steps

1. Open the Datasets panel and click Create a Dataset.

2. Name the dataset (for example, Anaplan - Events) and choose your Anaplan connection.

3. Select the Export Type that matches the dataset you are creating.

4. If prompted, choose the required Workspace and Model.

5. On the Define Columns screen, check IGNORE on every column the dataset does not need and set the types of the ones you keep. Each Export Type has its own list, given in full below and shown in that section’s screenshot. Do not skip this. The workflow template expects these exact column sets.

6. Save the dataset and repeat for each of the nine Export Types below.

---

Creating a Dataset

Every Anaplan audit dataset starts with the same three-step flow. Follow this pattern once to reach the dataset configuration screen, then use the Export Type specific settings shown below to finish each one.

Once your Data Connection is set to your Anaplan connection, you are ready to configure the Export Type.

---

Configure Each Export Type

After the connection is selected, choose the matching Export Type from the dropdown. Some types require additional Workspace or Model selections, and Events includes an interval setting. Every type also needs its columns set on the second screen, so work through both screenshots in each section below.

Anaplan - Events

Select Events as the Export Type and set Interval (hours) to 720, which pulls the trailing 30 days of audit events.

Choosing the Events Interval

Interval (hours) sets how far back each run reaches for audit events, so it is the one setting on this screen that depends on your own refresh strategy rather than on a fixed correct answer.

Lookback	Interval (hours)	When to use it
24 hours	24	Steady state, matched to a daily refresh
30 days	720	Initial load on standard Anaplan retention
90 days	2160	Initial load on Anaplan Enterprise, the maximum

Anaplan retains 30 days of audit events on standard editions and up to 90 days on Enterprise. Setting an interval beyond what your edition retains returns an error rather than a partial result. If you hit it, lower the value to 720 or 2160 to match your edition and run again.

Note

Load wide once, then narrow. A common approach is to set the interval to your maximum retention for the first run, 720 or 2160, so the audit log starts with real history behind it. Once that initial load is in place, drop the interval to match your refresh cadence, such as 24 for a daily refresh. Leaving it at the maximum means every run re-pulls the full window.

Use the built-in Nexadata scheduler to run the Workflow every 24 hours so the audit log keeps itself current without anyone starting it by hand.

Defining the Events Columns

Events is the one dataset where you keep everything, because the audit record itself is the data you are after. Leave every IGNORE box unchecked here.

Set the five numeric columns rather than leaving them as text:

Column	Type	Decimal Precision
id	Number	0
eventDate	Number	0
additionalAttributes.actionId	Number	2
checksum	Number	2
createdDate	Number	2

Everything else stays Text, apart from additionalAttributes.nux_visible and success, which are Boolean.

Anaplan - Users

Select Users as the Export Type to pull every user in the tenant.

On the Define Columns screen, keep id, userName, and displayName, all as Text. Ignore active, meta.created, meta.lastModified, meta.location, meta.resourceType, name.familyName, name.formatted, and name.givenName.

Note

Keep displayName. The template requires it and carries it through to the DISPLAY_NAME column in the final audit log. A Users dataset without it will not satisfy the template’s column requirement.

Anaplan - Workspaces

Select Workspaces as the Export Type to list every workspace in the tenant.

On the Define Columns screen, keep id and name, both as Text. Ignore active, currentSize, and sizeAllowance.

Anaplan - CloudWorks Jobs

Select CloudWorks Jobs as the Export Type to pull integration job records.

On the Define Columns screen, keep every column. Set latestRun.executionErrorCode to Number with a Decimal Precision of 0. latestRun.success and nuxVisible are Boolean, and everything else stays Text.

Anaplan - Models

Select Models as the Export Type, then choose the target Workspace.

On the Define Columns screen, keep id and name, both as Text. Ignore activeState, currentWorkspaceId, currentWorkspaceName, isoCreationDate, lastModified, lastModifiedByUserGuid, lastSavedSerialNumber, memoryUsage, and modelUrl.

---

One Set of Action Datasets Per Workspace and Model

The four Export Types that follow, Imports, Exports, Processes, and Actions, behave differently from the ones above. Each is scoped to a single Anaplan workspace and model rather than to the whole tenant, so a dataset built against one combination returns records for that combination only.

Note

Create all four datasets for every workspace and model combination you want in the audit output. A combination you leave out does not produce an error. Its events still appear in the audit log, but with no action name resolved against them, so activity is under-reported rather than visibly missing.

If you are covering more than one workspace and model combination, name each set so you can tell them apart later, for example Anaplan - Imports - Finance Prod. You will stack these additional datasets onto the Stack Actions supporting pipeline in Step 4.

Note

Reduce all four to the same four columns: id, modelId, name, and workspaceId. Each of these Export Types returns extra columns of its own, such as importType on Imports or exportFormat on Exports. On the Define Columns screen, check IGNORE on every column except id, modelId, name, and workspaceId. Set id to Number with a Decimal Precision of 0 and leave the other three as Text.

This is required, not cosmetic. The Stack Actions pipeline stacks these datasets on top of one another, which only works when they share an identical column structure, and the workflow template expects that structure to be in place when you apply it. A dataset that still carries its extra columns will not line up. modelId and workspaceId matter too. The template requires both, and the pipeline combines modelId with id to tell apart actions that share an ID across different models. Each section below lists exactly which columns to ignore.

Anaplan - Imports

Select Imports as the Export Type, then choose a Workspace and Model.

On the Define Columns screen, ignore importDataSourceId and importType, then set id to Number with a Decimal Precision of 0 and leave modelId, name, and workspaceId as Text.

Anaplan - Exports

Select Exports as the Export Type, then choose a Workspace and Model.

On the Define Columns screen, ignore encoding, exportFormat, exportType, and layout, then set id to Number with a Decimal Precision of 0 and leave modelId, name, and workspaceId as Text.

Anaplan - Processes

Select Processes as the Export Type, then choose a Workspace and Model.

On the Define Columns screen, Processes returns only id, modelId, name, and workspaceId, so there is nothing to ignore. Set id to Number with a Decimal Precision of 0 and leave the other three as Text.

Anaplan - Actions

Select Actions as the Export Type, then choose a Workspace and Model.

On the Define Columns screen, ignore actionType, then set id to Number with a Decimal Precision of 0 and leave modelId, name, and workspaceId as Text.

---

Back | Next