Problem reading data without local_cdi_id and edmo_code

[MartinHanssonSMHI]

We have problem reading ODV generated nc-files since the data do not contain local_cdi, edmo_code etc.

datafile2 = "C:/Work/DIVAnd/Test/data/data_from_Seabird_Collection_2022-06-01T13-55-00.nc"
@time obsval2, obslon2, obslat2, obsdepth2, obstime2, obsid2 = NCODV.load(Float64, datafile2, "Temperature");

┌ Warning: No variable with the long_name attribute 'LOCAL_CDI_ID' in C:/Work/DIVAnd/Test/data/data_from_Seabird_Collection_2022-06-01T13-55-00.nc found. We use the empty string for LOCAL_CDI_ID instead.
└ @ DIVAnd.NCODV C:\Users\a001109.julia\packages\DIVAnd\Iwjb1\src\NCODV.jl:327
No NetCDF variable found with attribute 'long_name' equal to 'EDMO_CODE' in C:\Work\DIVAnd\Test\data\data_from_Seabird_Collection_2022-06-01T13-55-00.nc

Stacktrace:
[1] error(s::String)
@ Base .\error.jl:33
[2] varbyattrib_first(ds::NCDataset{Nothing}; kwargs::Base.Iterators.Pairs{Symbol, String, Tuple{Symbol}, NamedTuple{(:long_name,), Tuple{String}}})
@ DIVAnd.NCODV C:\Users\a001109.julia\packages\DIVAnd\Iwjb1\src\NCODV.jl:15
[3] (::DIVAnd.NCODV.var"#10#11"{Int64, DataType, String, String, Vector{String}})(ds::NCDataset{Nothing})
@ DIVAnd.NCODV C:\Users\a001109.julia\packages\DIVAnd\Iwjb1\src\NCODV.jl:348
[4] NCDataset(f::DIVAnd.NCODV.var"#10#11"{Int64, DataType, String, String, Vector{String}}, args::String; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
@ NCDatasets C:\Users\a001109.julia\packages\NCDatasets\XVX8L\src\dataset.jl:220
[5] NCDataset(f::Function, args::String)
@ NCDatasets C:\Users\a001109.julia\packages\NCDatasets\XVX8L\src\dataset.jl:218
[6] #load#9
@ C:\Users\a001109.julia\packages\DIVAnd\Iwjb1\src\NCODV.jl:322 [inlined]
[7] load(T::Type, fname::String, long_name::String)
@ DIVAnd.NCODV C:\Users\a001109.julia\packages\DIVAnd\Iwjb1\src\NCODV.jl:320
[8] top-level scope
@ .\timing.jl:210 [inlined]
[9] top-level scope
@ .\In[11]:0
[10] eval
@ .\boot.jl:360 [inlined]
[11] include_string(mapexpr::typeof(REPL.softscope), mod::Module, code::String, filename::String)
@ Base .\loading.jl:1116

Julia 1.6.6

DIVAnd.jl/src/NCODV.jl

Line 345 in fbf6867

EDMO_CODE = if length(varbyattrib(ds; long_name = "EDMO_code")) > 0

[Alexander-Barth]

Even when by-passing the check of EDMO_CODE, I get an error when loading depth.
How come that you variable depth variable is not called "Depth" ? I only see these:

      float var20(N_STATIONS, N_SAMPLES) ;
                var20:long_name = "Depth [true depth, m], lat = 58.0168" ;
                var20:units = "" ;
                var20:comment = "" ;
                var20:ancillary_variables = "var20_qc" ;
                var20:C_format = "%.2f" ;
                var20:FORTRAN_format = "F12.2" ;
                var20:_FillValue = -1.e+10f ;

       float var19(N_STATIONS, N_SAMPLES) ;
                var19:long_name = "Depth [salt water, m], lat = 58.0168" ;
                var19:units = "" ;
                var19:comment = "" ;
                var19:ancillary_variables = "var19_qc" ;
                var19:C_format = "%.2f" ;
                var19:FORTRAN_format = "F12.2" ;
                var19:_FillValue = -1.e+10f ;

[MartinHanssonSMHI]

Its generated in the SBE-software while processing and can be remove. We usually use pressure and can rename pressure to depth. MVH/M

…

_________________________________________________________________________ Martin Hansson Oceanographic Unit Swedish Meteorological and Hydrological Institute (SMHI) National Oceanographic Data Centre (NODC) Sven Källfelts gata 15 SE-421 71 Västra Frölunda Sweden Phone: +46 (0) 31 7518957 Mob: +46 (0) 700 900012 www.smhi.se/tema/havsmiljo<http://www.smhi.se/tema/havsmiljo> Från: Alexander Barth ***@***.***> Skickat: den 16 september 2022 14:04 Till: gher-uliege/DIVAnd.jl ***@***.***> Kopia: Hansson Martin ***@***.***>; Author ***@***.***> Ämne: Re: [gher-uliege/DIVAnd.jl] Problem reading data without local_cdi_id and edmo_code (Issue #106) How come that you variable depth variable is not called "Depth" ? I only see these: float var20(N_STATIONS, N_SAMPLES) ; var20:long_name = "Depth [true depth, m], lat = 58.0168" ; var20:units = "" ; var20:comment = "" ; var20:ancillary_variables = "var20_qc" ; var20:C_format = "%.2f" ; var20:FORTRAN_format = "F12.2" ; var20:_FillValue = -1.e+10f ; float var19(N_STATIONS, N_SAMPLES) ; var19:long_name = "Depth [salt water, m], lat = 58.0168" ; var19:units = "" ; var19:comment = "" ; var19:ancillary_variables = "var19_qc" ; var19:C_format = "%.2f" ; var19:FORTRAN_format = "F12.2" ; var19:_FillValue = -1.e+10f ; — Reply to this email directly, view it on GitHub<#106 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AT7EQN2LBGKN4UYHI7AS2ULV6RO3HANCNFSM6AAAAAAQOH6WV4>. You are receiving this because you authored the thread.Message ID: ***@***.******@***.***>>

[jmbeckers]

I think in the old DIVA version we searched for depth and if not found used pressure instead. Would that be a solution or is it better to force DIVAnd users to stick to depth ?

[Alexander-Barth]

Well depth is there, but it is not called in a way that we expected it in DIVAnd. Unfortunately, ODV does not use the CF standard names (nor the BODC names) so we have to rely on long_name, but this assumes that the long_name is consistently named (Depth as in the example files that I got from Reiner, but not Depth [true depth, m], lat = 58.0168). This should be easy that handle in ODV (outside of DIVAnd).

Please feel free to re-open if you have trouble renaming the attribute or ask in the ODV forum.