We have noticed that for many of our collections, irrelevant dimension scale variables are included in variable subsets. E.g. see the following request for S5P_L2__HCHO___HiR_2
https://harmony.earthdata.nasa.gov/C1918210023-GES_DISC/ogc-api-coverages/1.0.0/collections/parameter_vars/coverage/rangeset?forceAsync=true&subset=lat(0%3A50)&subset=lon(50%3A110)&subset=time(%222025-11-23T00%3A00%3A00%22%3A%222025-12-31T00%3A00%3A00%22)&maxResults=1&variable=PRODUCT%2Fformaldehyde_tropospheric_vertical_column&variable=PRODUCT%2Fqa_value&variable=PRODUCT%2Ftime_utc&variable=PRODUCT%2FSUPPORT_DATA%2FGEOLOCATIONS%2Flatitude_bounds&variable=PRODUCT%2FSUPPORT_DATA%2FGEOLOCATIONS%2Flongitude_bounds&variable=PRODUCT%2FSUPPORT_DATA%2FINPUT_DATA%2Fsurface_pressure
with the following request we get at least 10 variables that are not requested nor are used for the requested variables. 6 are in the METADATA/QA_STATISTICS group - ground_pixel, histogram_axis, pdf_axis, scanline, time, and vertices. 4 are un /PRODUCT/SUPPORT_DATA/INPUT_DATA/BACKGROUND_CORRECTION - detector_rows, latitude_grids, polynomial_coefficients, and wavelengths. None of the requested variables actually use these dimension scale variables and for the QA_STATISTICS group the submitter is actually /adding/ 3 variables that did not previously exist in that group.
I have a solution working and will open a PR to address this in the next day or so.
We have noticed that for many of our collections, irrelevant dimension scale variables are included in variable subsets. E.g. see the following request for
S5P_L2__HCHO___HiR_2with the following request we get at least 10 variables that are not requested nor are used for the requested variables. 6 are in the
METADATA/QA_STATISTICSgroup -ground_pixel,histogram_axis,pdf_axis,scanline,time, andvertices. 4 are un/PRODUCT/SUPPORT_DATA/INPUT_DATA/BACKGROUND_CORRECTION-detector_rows,latitude_grids,polynomial_coefficients, andwavelengths. None of the requested variables actually use these dimension scale variables and for theQA_STATISTICSgroup the submitter is actually /adding/ 3 variables that did not previously exist in that group.I have a solution working and will open a PR to address this in the next day or so.