Allow multi axes queries in execute_mdx function

[MariusWirtz]

The new change allows this:

from TM1py import TM1Service

with TM1Service(base_url="https://localhost:12354", user="admin", password="apple") as tm1:
    mdx = """
    SELECT
    NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d1])},0) on 0,
    NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d2])},0) on 1,
    NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d3])},0) on 2,
    NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d4])},0) on 3,
    NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d5])},0) on 4
    FROM [c5]
    """

    cells = tm1.cells.execute_mdx(mdx, element_unique_names=False, skip_cell_properties=True)
    for elements, value in cells.items():
        print(elements, value)

('e1', 'e1', 'e01', 'e01', 'e01') 4
('e1', 'e1', 'e01', 'e02', 'e01') 1
('e1', 'e1', 'e01', 'e01', 'e02') 2
('e1', 'e1', 'e01', 'e02', 'e02') 3

[wimgielis]

Hello Marius, For me in a test cube, PAW sends axes 2, 3, 4 to the bench and only shows 0 and 1. Is this expected behaviour or is this MDX view only used for getting values instead of an object for the users in the interfaces ? Best regards, Wim

…

------ Best regards / Beste groeten, Wim Gielis MS Excel MVP 2011-2014 https://www.wimgielis.com <http://www.wimgielis.be> Op ma 8 aug. 2022 om 15:37 schreef Marius Wirtz ***@***.***>:

The new change allows this: from TM1py import TM1Service with TM1Service(base_url="https://localhost:12354", user="admin", password="apple") as tm1: mdx = """ SELECT NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d1])},0) on 0, NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d2])},0) on 1, NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d3])},0) on 2, NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d4])},0) on 3, NON EMPTY Tm1FilterByLevel({TM1SubsetAll([d5])},0) on 4 FROM [c5] """ cells = tm1.cells.execute_mdx(mdx, element_unique_names=False, skip_cell_properties=True) for elements, value in cells.items(): print(elements, value) ('e1', 'e1', 'e01', 'e01', 'e01') 4 ('e1', 'e1', 'e01', 'e02', 'e01') 1 ('e1', 'e1', 'e01', 'e01', 'e02') 2 ('e1', 'e1', 'e01', 'e02', 'e02') 3 ------------------------------ You can view, comment on, or merge this pull request online at: #777 Commit Summary - e43d64f <e43d64f> Allow multi axes queries in `execute_mdx` function File Changes (2 files <https://github.com/cubewise-code/tm1py/pull/777/files>) - *M* TM1py/Utils/Utils.py <https://github.com/cubewise-code/tm1py/pull/777/files#diff-dfa6efa267891cf338ac4ae2fb7d41b941f6ecef55b265005ff460b33c9e8f9d> (42) - *M* Tests/CellService_test.py <https://github.com/cubewise-code/tm1py/pull/777/files#diff-d044783f3dd2a3bc2510c242076847e316dd0ce57d3724cc09870fcca88f5e2d> (59) Patch Links: - https://github.com/cubewise-code/tm1py/pull/777.patch - https://github.com/cubewise-code/tm1py/pull/777.diff — Reply to this email directly, view it on GitHub <#777>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEDHULP2TZPVXPPSDCAFVTTVYEEQBANCNFSM555FIAFA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[rclapp]

I'd say PAW doesn't know what to do with additional axis. I'm not sure there is any Standard tbh. I would like PAW to reflect multiple axis with something like tabs or pages.

[MariusWirtz]

@wimgielis I don't know what PAW does in detail, but the API responses are consistent with the MDX queries.

We noticed that when using asymmetric tuples in the selections, multi-axes queries perform much better than queries with stacked selections on 0 or 1.

[rclapp]

@MariusWirtz do you have an example query that you can share?

[MariusWirtz]

@rclapp here is a sample:

SELECT
NON EMPTY {
([Cost Center].[10], [Company].[22]),
([Cost Center].[29], [Company].[62]),
([Cost Center].[15], [Company].[23])
} ON 0,
NON EMPTY {
([Year].[2022], [Month].[1]),
([Year].[2021], [Month].[2]),
([Year].[2020], [Month].[3])
} ON 1,
NON EMPTY {
([State].[CA], [Version].[Actual]),
([State].[OK], [Version].[Budget]),
([State].[FL], [Version].[Forecast]),
} ON 2,
NON EMPTY {
([PC].[110], [LOB].[55]),
([PC].[210], [LOB].[33]),
([PC].[92], [LOB].[12]),
} ON 3
FROM [GL]

The above executes faster than the below in my tests.

SELECT
NON EMPTY {
([Cost Center].[10], [Company].[22]),
([Cost Center].[29], [Company].[62]),
([Cost Center].[15], [Company].[23])
} ON 0,
NON EMPTY {
([Year].[2022], [Month].[1]),
([Year].[2021], [Month].[2]),
([Year].[2020], [Month].[3])
} * {
([State].[CA], [Version].[Actual]),
([State].[OK], [Version].[Budget]),
([State].[FL], [Version].[Forecast]),
} * {
([PC].[110], [LOB].[55]),
([PC].[210], [LOB].[33]),
([PC].[92], [LOB].[12]),
} ON 1
FROM [GL]