UNST-9376: Accept both "waterlevel" and "waterlevelbnd" #369
Conversation
ec_typedefs::tEcNetCDF: Register what variables are time varying ec_netcdf_timeseries.f90::ecNetCDFInit: - Fill array time_varying_var - Vertical layering should be identified with "axis=Z" instead of "positive = notEmpty" - Vertical layering: accept standard_names "z, zcoordinate, zcoordinate_c", exclude "zcoordinate_w, zcoordinate_wu" ec_netcdf_timeseries.f90::ecNetCDFScan: - ltl is defined by calling len_trim_nobnd ec_netcdf_timeseries.f90::len_trim_nobnd: - New function, returning the length of a string excluding "bnd" at the end
|
@arthurvd : At some point we need to discuss how to facilitate 3D quantities for nesting, including z-coordinates varying in time. This PR is a first step, enabling Theo to continue. |
When reading "zcoordinate_c"-related info, the nesting functionality is ruined. TODO: find out why and implement a proper fix.
| if (ncptr%vptyp < 1) then | ||
| call setECMessage("ec_bcreader::ecNetCDFCreate: Unable to determine vertical coordinate system.") | ||
| ! This line does not work: if (any((/ 'z ', 'zcoordinate ', 'zcoordinate_c'/) == ncptr%standard_names(iVars))) then | ||
| if (any((/ 'z ', 'zcoordinate '/) == ncptr%standard_names(iVars))) then |
There was a problem hiding this comment.
Maybe we need to do this differently, let's refine a proper 3D initialization under the other dedicated issue: UNST-8909 (so not do this here)
| if (ncptr%vptyp < 1) then | ||
| call setECMessage("ec_bcreader::ecNetCDFCreate: Unable to determine vertical coordinate system.") | ||
| ! This line does not work: if (any((/ 'z ', 'zcoordinate ', 'zcoordinate_c'/) == ncptr%standard_names(iVars))) then | ||
| if (any((/ 'z ', 'zcoordinate '/) == ncptr%standard_names(iVars))) then |
There was a problem hiding this comment.
p.s. Coding guidelines require the modern format for initialization arrays, namely: [ .. ] instead of (/ .. /).
| if (ncptr%nLayer > size(ncptr%vp)) then | ||
| call realloc (ncptr%vp, ncptr%nLayer, stat=ierr, keepExisting=.true.) | ||
| end if | ||
| if (ierr /= 0) return |
There was a problem hiding this comment.
If 3D work is postponed until UNST-8909, why not simply have an if .. cycle here?
| if (ncptr%nLayer > size(ncptr%vp)) then | |
| call realloc (ncptr%vp, ncptr%nLayer, stat=ierr, keepExisting=.true.) | |
| end if | |
| if (ierr /= 0) return | |
| allocate (ncptr%vp(ncptr%nLayer), stat=ierr) | |
| if (ierr /= 0) cycle |
| do ivar = 1, ncptr%nVars | ||
| ltl = len_trim(quantity) | ||
| if (strcmpi(trim(ncptr%standard_names(ivar)), trim(quantity))) exit | ||
| if (strcmpi(trim(ncptr%standard_names(ivar)), trim(quantity), ltl)) exit |
There was a problem hiding this comment.
I would hope we can make this work instead using the existing ncstdnames approach, by having something like this in ecProviderCreateNetCdfItems():
case ('waterlevelbnd')
ncvarnames(1) = 'waterlevel'
ncstdnames(1) = 'sea_surface_height'
Let's refine this together under UNST-9376 for proper followup in next sprint.
Commit Description
ec_typedefs::tEcNetCDF: Register what variables are time varying
ec_netcdf_timeseries.f90::ecNetCDFInit:
ec_netcdf_timeseries.f90::ecNetCDFScan:
ec_netcdf_timeseries.f90::len_trim_nobnd:
What was done
TODO
Evidence of the work done
<add video/figures if applicable>
Tests
<add testcase numbers if applicable, Issue number>
Documentation
<add description of changes if applicable, Issue number>
Issue link
UNST-9376