Cube must be 2-dimensional. Got 3 dimensions; Help to edit the diag script

[AnushaNTNU]

Hi @rswamina and All
I just changed the input files in recipe_python.yml like below
datasets:

• {dataset: BCC-ESM1, project: CMIP6, exp: historical, ensemble: r1i1p1f1, grid: gn}
• {dataset: CanESM2, project: CMIP5, exp: historical, ensemble: r1i1p1}
  datasets:
• {dataset: NorESM2-LM, project: CMIP6, exp: ssp534-over-bgc, ensemble: r1i1p1f1, grid: gn}
• {dataset: CanESM5, project: CMIP6, exp: ssp534-over-bgc, ensemble: r1i1p2f1, grid: gn}
  And I get the error
  ValueError: Cube must be 2-dimensional. Got 3 dimensions. can you help me to edit the diag script?
  Attaching the run/recipe_*.yml and run/main_log_debug.txt from the output directory
  main_log_debug.txt
  recipe_python_AS.yml.txt

[rswamina]

@AnushaNTNU - Can you share the following file examples/diagnostic.py failed with return code 1. See the log in /projects/NS9576K/www/diagnostics/esmvaltool/a2021/tmp/recipe_python_AS_20221011_132922/run/map/script1/log.txt?

[AnushaNTNU]

Yes sure. Attaching it here.
log.txt

[rswamina]

@AnushaNTNU - In this case, the plot routine gets a 3D cube instead of 2D. In the original recipe, a single month from a single year is selected so it is 2D. See below

        preprocessor: select_january
        start_year: 2000
        end_year: 2000

In your recipe, you have selected more than one year so , you will have a time dimension with length 60. In the diagnostic, you might have to calculate average over time or add a time mean operator in the preprocessor.

        preprocessor: select_january
        start_year: 2040
        end_year: 2100

Hope that helps.

[AnushaNTNU]

Yes . Thanks a lot, its clear now. But changing the recipe or diag script is still not easy for me. I could just succesfully remove the timeseries and map for the particular location (Amsterdam) from recipe so that it plots only global mean timeseries for january from 2040 to 2100 and map for january 2100. Can I write annual_mean_global for processor instead of select_january?

[rswamina]

Yes. You can, if that's what you want. have you had a chance to look at th tutorial yet? You could also use preprocessors and code from diagnostic in other recipes. Depends on what you would like. Let us know if you have trouble.

[AnushaNTNU]

Yes, I am learning from it. But I feel its insufficient. Modifying the preprocessor is a bit confusing. I will watch the videos soon and lets see. Thanks for the help