TM1py 1.6

[MariusWirtz]

Hello TM1 community,

Last week we released TM1py 1.6.
It's a big release with interesting new features like impersonation and improved write speed.

You can find an overview of the release here:
https://code.cubewise.com/blog/tm1py-v1 ... -available

Are you using TM1py? If not, may I ask why?

[jim wood]

Post moved. Please read the forum guidelines for commercial related posts like this going forward.

[Paul Segal]

Slightly harsh, Jim. If we were talking about arc or pulse then fair enough, but tm1py seems to me to be the epitome of open source and non-commercial.

Regards

[jim wood]

Slightly harsh, Jim. If we were talking about arc or pulse then fair enough, but tm1py seems to me to be the epitome of open source and non-commercial.

Regards

I hear you, but we try to keep product releases (open source or otherwise) out of the main forum as things start to get very shaky very quickly. Some open source products / services are used as marketing tools, granted not all. But if you start making an exception for one then you put other people out who want theirs etc etc.

[scrumthing]

Very cool release.
I especially love the write_dataframe method after some initial struggle with our German locale on the server side.
it makes some of the fancy data science **[Expletive Removed]** you can do with Python a lot easier to get back into the cube.

The windows authentication is awesome. Just that makes the direct connection to tm1 sooooo easy compared to getting SSO up and running for TM1Web. Just that would be a reason to switch to PowerBI as a TM1 frontend for smaller environments.

[lotsaram]

I hear you, but we try to keep product releases (open source or otherwise) out of the main forum as things start to get very shaky very quickly. Some open source products / services are used as marketing tools, granted not all. But if you start making an exception for one then you put other people out who want theirs etc etc.

FWIW tm1py is genuinely open source. Anyone can access the code, anyone can contribute, anyone can use it (for free).

I don't understand the motivation for moving the post, but if you're going to invoke the slipery slope argument then be consistent. By the exact reasoning you mention then bgregg's Monaco editor or Notepad++ plugin or anytime anyone posts a link to a git repository or asks for input or feedback on a personal side project then this is also going to come under your definition.

[Emixam]

Are you using TM1py?

Yes I do! I'm currently working on a LSTM network (deep learning) and I'm using TM1py to pull data from a cube and will eventually send my results back into TM1. Unfortunately, data weren't clean enough so we need to work on our sources but I'm excited to share that with the community when I'll be done.

TM1py is a game changer for people like me, thanks!

[michaelc99]

Hello TM1 community,

Last week we released TM1py 1.6.
It's a big release with interesting new features like impersonation and improved write speed.

You can find an overview of the release here:
https://code.cubewise.com/blog/tm1py-v1 ... -available

Are you using TM1py? If not, may I ask why?

Good Morning Marius,

I just watched a video clip from a TM1 User Conference in Sydney, Australia where you and Vincent Viau were discussing TM1 Rest API and TM1py. I must say that I am impressed with the capabilities and I'm excited to begin development on cube enhancements.

One challenge that I'm preparing to face (implementation is about 9-12 months out) is with respect to connecting to a Power BI Premium Dataset. We want the capability to 1) Load monthly data from Power BI into TM1 (though, my preference is to go from DAX365 directly) and 2) Build a Drill-Through capability from TM1 into Power BI (Premium Dataset) for transactional level details.

I'm assuming based on the above comments that both are achievable, but I was hoping that you (or someone else) could confirm my suspicions to be true.

Thank you in advance,
Michael

[MariusWirtz]

Hi Michael,

thanks!

Am I understanding you right, that your desired data flow is PowerBI -> TM1?
Interesting!

Once you have the data in python it is straightforward to bring it to a TM1 cube with TM1py.
However, the first part for python to retrieve data from a PBI model could be a bit more challenging.
In the PBI forum, you can probably get a more precise answer on how this can be accomplished.
Someone actually posted the same question in the PBI forum not long ago:
https://community.powerbi.com/t5/Develo ... -p/1352246

A native TM1 drill-through (like you do with SQL or CSV data sources) is not possible with PBI as a data source for a cube.
I can think of two alternatives:

1. You build a kind of staging / middle layer in SQL.
From there data is written to the cube, and users can use drill-through capabilities to drill from a cube cell into this staging layer.

2. When a drill-through is executed by a user, python retrieves the requested slice from PBI on the fly.
Python writes the data into an identifiable flat file or a table and then the TI process drills into that flat file or table.

[michaelc99]

Hi Michael,

thanks!

Am I understanding you right, that your desired data flow is PowerBI -> TM1?
Interesting!

Once you have the data in python it is straightforward to bring it to a TM1 cube with TM1py.
However, the first part for python to retrieve data from a PBI model could be a bit more challenging.
In the PBI forum, you can probably get a more precise answer on how this can be accomplished.
Someone actually posted the same question in the PBI forum not long ago:
https://community.powerbi.com/t5/Develo ... -p/1352246

A native TM1 drill-through (like you do with SQL or CSV data sources) is not possible with PBI as a data source for a cube.
I can think of two alternatives:

1. You build a kind of staging / middle layer in SQL.
From there data is written to the cube, and users can use drill-through capabilities to drill from a cube cell into this staging layer.

2. When a drill-through is executed by a user, python retrieves the requested slice from PBI on the fly.
Python writes the data into an identifiable flat file or a table and then the TI process drills into that flat file or table.

Hello Marius,

Thank you for responding.

We are currently on MS Dynamics AX 2012, but are moving to MS Dynamics 365 in about 6-9 months. Previously, our process was to export Excel files from AX and load them from Excel into TM1; however, I built VBA code to covert the Excel files into CSV, so that TI processes can load data directly into TM1. There are manual processes still, but I will continue finding ways to eliminate them.

When we move to D365, I will look to write TI processes to automatically export and load the data into TM1, which is my preferred approach. If I run into any complications, then I may need to either go back to the manual approach (above, not ideal) or pull data from a Power BI premium dataset. I am also looking to create a drill-through process to go from TM1 to a Power BI premium dataset for transaction level details.

I will look into a staging/middle layer in SQL and see what I can do there and I will also checkout that link as well.

Thank you!