Open elray1 opened 4 months ago
In general, it seems like it's hard to deal with this because the output_type_id
could be numeric or character and we don't know which way to treat it. Maybe one way to deal with this is to assert that the hub has listed the values in their tasks.json
file in the correct order, and do the sorting here in the same order as they were given there? Some care would need to be taken to split by round, in case different levels were used in the different rounds...
The helper function
check_values_ascending
arranges by theoutput_type_id
here, which can result in an incorrect ordering and an incorrect validation failure when theoutput_type_id
has a character data type but represents numeric values. This hasn't come up in testing with quantile outputs because theoutput_type_id
s for those have a leading 0 -- but that may not be the case for cdf output types.Here is an example:
If it's more helpful to have an example in data file format, the example above is a subset of this file.