Wrong Pivot Table entries are expanded in QlikView 12 with Reduce Data
Article Number: 000034186 | Last Modified: 2018/06/12
In QlikView 12, when user collapse and expand some nodes in a pivot table, make some selection and reduce data using function File > Reduce Data > Keep Possible Values, the "collapse-expand" layout of the nodes is then displayed differently in the reduced qvw document. Some nodes that were previously collapsed are now expanded and vice versa.
The behaviour is the same with reduction option for reload and distribution tasks in QlikView Publisher. This causes problems when QlikView admin needs to reduce data and distribute PDF reports containing pivot tables to end-users.
The problem did not appear in QlikView 11.20
This comes from new implementation and optimizations in QlikView 12 and it is considered as WAD.
This is considered as WAD.
As reported by the development team, the correct way to secure a consistent behaviour regarding collapse/expand status of a dimension in a pivot table, is to base the collapse/expand information on the load order of the values in the dimension.
Therefore, when the load order is random a consistent load order needs to be established in the script.
There are two possible workarounds.
The first workaround would be to establish a well-defined sort order of all relevant fields in the script, making sure that the Load-order of a filed is the desired and intended order. This may or may not be a complex operation due to the number of fields involved and in this particular case, these fields are populating the pivot table in a fairly dynamic way, so it may be a relatively complicated operation.
So you would have to create the sort order for the fields in the Pivot table manually in the script.
Another possibility would be to use the Presentation tab setting: "Always Fully Expanded", but even this setting may be subject to an override if both the list of dimensions and the sort order within relevant dimensions are affected by the reload.