Multi threading in planning analytics

[pandeytm777]

Hi,
I am trying to find more information related to multi-threading in Planning analytics. Did not find any great examples to implement this in TM1. I want to load the data from source (csv file with 500,000 records) to a staging cube in TM1 using multi-threading feature to improve performance. Any suggestions or samples around multi-threading would be highly appreciated. I am using PA 2.0.9.

Thank you!

[tomok]

Hi,
I am trying to find more information related to multi-threading in Planning analytics. Did not find any great examples to implement this in TM1. I want to load the data from source (csv file with 500,000 records) to a staging cube in TM1 using multi-threading feature to improve performance. Any suggestions or samples around multi-threading would be highly appreciated. I am using PA 2.0.9.

Thank you!

How hard did you try? I typed in "MTQ in TM1" into Google search box and pulled back a number of links from IBM documentation fully explaining how it works.

[ByronB]

Hi @pandeytm777,

About Parallel loading, I would recommend you to read this article (https://code.cubewise.com/blog/4-ways-t ... -analytics). It explains 4 different ways to do parallel loading in TM1.

Tomok, MTQ and parallel loading are two separate things. Running a TM1 process might not always trigger MTQ. To learn more about MTQ: https://code.cubewise.com/blog/masterin ... -analytics

Cheers,

[pandeytm777]

Thank you ByronB and tomok for your responses. I did find these articles. Yes, parallel processing and multi-threading are different. Multi-threading applies while reading or processing the data from the source. Parallel processing applies to loading of the data into TM1 to increase the processing time. I am mostly interested in the multi-threading aspect while reading the data. I believe multi-threading basically involves setting the different config paramters in the tm1s.cfg file. As you mentioned, TM1 process may not trigger MTQ in every case. My confusion is around the same thing. How do I force tm1 TI process to use multi-threading in my scenario while reading the data from the csv file. Is it possible? Are there any TI commands to enforce MTQ? My file has month dimension. So, is it possible to have one thread for each month while reading the data and then load it using parallel processing(tm1runTI)?

[lotsaram]

I am mostly interested in the multi-threading aspect while reading the data. I believe multi-threading basically involves setting the different config paramters in the tm1s.cfg file. As you mentioned, TM1 process may not trigger MTQ in every case. My confusion is around the same thing. How do I force tm1 TI process to use multi-threading in my scenario while reading the data from the csv file. Is it possible?

What you are expecting exceeds the current capabilities of Planning Analytics. If you want to use multiple threads to load a large external file then you need to first split the file and then load each file in a separate thread (or for example if your source was ODBC then use multiple smaller queries by month not one big query of the whole year).