Import Saved View - Filter

MaartenGijbels

Hi all,

 

A bit surprised this topic isn't here as it's one of the most common frustrations, similar to the post here. When importing from a filtered saved view there are a lot of cases of missing values.

 

We all encounter it, and the first time it takes very long to determine the root cause. We change the filters but then the view becomes bigger which impacts runtime of the action.

 

When a filter is applied on a saved view with page selectors it usually creates issues if other cells are populated for other page selectors. For exports there is a solution by filtering on the way out:

 

 

 

It would be logical that a saved view that is being imported is filtered in the same way. The filter is applied for all possible combinations of the page selectors. Not for the current selections and "enforced" on the other combinations.

 

All the best,

Maarten.

Miran

Status changed to: New

david.savarin

Completely support this.

We ran into this issue previously and the workaround is to remove the filters from the saved view with the following side effects :

• Import actions take a lot of time to process
• Now we get error messages that we cannot import more than 2^31 cells

GregA

This is intended behaviour. Filtering is for grid views, and pages take you between different grid views (with data split by, e.g, Account). Good for more user-facing models, but less so for data manipulation.

The solution is to ensure that you are importing (ideally from a data hub) from a view which has all the data you need on ONE grid, filtered (e.g to avoid importing thousands of zero values).

So if you have pages for Accounts, and rows for SKUs, pivot the saved view so that you have one big grid of ALL Accounts, with data subsets split by SKU.

You import the same reduced dataspace, without the Anaplan engine having to think 'Do they want me to import over just this page? Or across all the pages in this saved view?'

Hope this helps anyone wanting to make use of importing through saved views!

david.savarin

What if we have like 5 or 6 dimensions ? We cannot pivot all dimensions in lines so we are stuck with some heavy workarounds

GregA

Filtering should not be a primary line of defence in maintaining structure for your models. If, for instance, you have a module which has

• Number of customer interactions
• Length of Customer Interaction

line items, split by

• Location
• Department
• Customer Subscription Type
• Date,

you would of course be able to maintain some order by using filters (e.g, show only customer interaction lengths above 50 from Boston's Sales team of Platinum subscription type), but you would not be able to import your data reliably across all the different cross-sections you might like. You would also slow the performance of your model to a crawl, because you're importing ALL cells, many of which are unnecessary for your analysis.

As an aside, doing things this way also makes the model hard to understand and audit. The issue becomes less about 'why doesn't importing with filters recalculate across all the page dimensions' and more to do with general model optimisation.

https://community.anaplan.com/t5/Best-Practices/PLANS-This-Is-How-We-Model/ta-p/33530