Change in behaviour updating attributes across different versions of PA

[rockysystems]

Hi ,

I have a script that's based on an example for re-syncing attributes and which has been running great in my local environment
https://cubewise-code.github.io/tm1py-tales/2021/sync-dimensions.html

However, when I run the code against a more recent version of PA the code is crashing with an error of
Element "Base" not found in dimension - base is neither an element or attribute

The dimension in question contains hierarchies so base feels like the primary/leaf hierarchy but only occurs when I run against latter versions of PA.

If I output the cellset I can see a difference between the versions of PA, for example

PAL v11.8.02300.10 (this works)
100001,Group
100001,Band

PAL v11.8.03200.3 - note the addition of Base
Base,100001,Group
Base,100001,Band

So this would imply there's been a change in behaviour somewhere down the line. Just to be sure, I have also upgraded TM1PY to 2.2 but no change.

Has anyone come across anything similar? or any thoughts? At the moment I can't get the code to work for the latest versions of PA

Many thanks,
Steven

Extract of my code below
`

def promote_attributes(self, dimension_name):
    attribute_cube = "}ElementAttributes_" + dimension_name

    query = MdxBuilder.from_cube(attribute_cube)
    query = query.non_empty(axis=0)
    query = query.add_hierarchy_set_to_column_axis(MdxHierarchySet.tm1_subset_all(dimension_name))
    query = query.add_hierarchy_set_to_column_axis(MdxHierarchySet.tm1_subset_all(attribute_cube))
    mdx = query.to_mdx()

    attribute_cells = self.source.cells.execute_mdx(
        mdx=mdx,
        element_unique_names=False,
        skip_cell_properties=True,
        skip_rule_derived_cells=True)


    clear_process = Process(f"CubeClearData('{attribute_cube}');")
    success, _, _ = self.target.processes.execute_process_with_return(clear_process)
    if not success:
        raise RuntimeError(f"Failed to clear attribute cube for dimension: '{dimension_name}'")

    # Check if async is enabled in the config file
    # async allows multi-threaded loads based on slice size and max_workers and should support faster load times
    if self.tm1config['options']['async']:
        self.target.cells.write_async(
            cube_name=attribute_cube,
            cells=attribute_cells,
            slice_size=self.tm1config['options']['slice_size'],
            max_workers=self.tm1config['options']['max_workers'],
            increment=False,
            deactivate_transaction_log=True,
            reactivate_transaction_log=True)
    else:
        self.target.cells.write(
            cube_name=attribute_cube,
            cellset_as_dict=attribute_cells,
            use_ti=True,
            deactivate_transaction_log=True,
            reactivate_transaction_log=True)

    self.logger.info(f"Attributes resynced for '{dimension_name}'.")`

[MariusWirtz]

Perhaps this is related to the EnableSandboxDimension config parameter?
https://www.ibm.com/docs/en/planning-analytics/2.0.0?topic=parameters-enablesandboxdimension

Does the response change if you set it to False?

[rockysystems]

Brilliant, thanks Marius, it was indeed and I think the last thing I would have checked ..
I've also added skip_sandbox_dimension=False which resolved the problem.