Data importer - parsing units should be case sensitive

[PavelBal]

A while ago I created an issue that parsing of units during import should be case-insensitive #1949

Unfortunately, I have a case where this leads to wrong dimension - when trying to import µm of the dimension Lenght, the importer regodnizes it as µM of the dimension Concentration (molar).

[msevestre]

well that's ok. We cannot have it both ways. So when you import, it's your job as a user to verify that units and dimensions are recognized properly.

[PavelBal]

Yes but then I would vote for having it case-sensitive. Otherwise there is no way to import anything other than dm for length without having to manually change the dimension (== no way to automatize it through R)

[Yuri05]

use µmol/L instead

[Yuri05]

The most flexible way:

1. Try to find the unit case sensitive first. If found: use it
2. If not found: try to find the unit case insensitive

[msevestre]

I think this is the way it's actually implemented. First case sensitive, then case insensitive

[Yuri05]

I think this is the way it's actually implemented. First case sensitive, then case insensitive

But then µm should be recognized as µm and not as µM
@PavelBal Can you describe the steps to reproduce you problem?

[georgeDaskalakis]

It is not an issue of unit recognition but of automatically detecting the correct dimension. So (if I am not mistaken by what Pavel describes), in the calculation of the dimension from a unit given in an excel file:

https://github.com/Open-Systems-Pharmacology/OSPSuite.Core/blob/50a7aa040611e25be40541b8a8748976946808a7/src/OSPSuite.Infrastructure.Import/Core/DataFormat/AbstractColumnsDataFormat.cs#LL159C1-L167C8

we are not trying case sensitive first, we just get the first dimension that has such a unit. In Pavel's case with µm , the first dimension found that has such a unit ( µM in this case, since case insensitive) would be Concentration (molar).

[georgeDaskalakis]

I think this is the way it's actually implemented. First case sensitive, then case insensitive

Actually implementing this would be a fix that would keep the correct behaviour also for #1949

[Yuri05]

It is not an issue of unit recognition but of automatically detecting the correct dimension. So (if I am not mistaken by what Pavel describes), in the calculation of the dimension from a unit given in an excel file:

I think it would be better to get the dimension from unit via DimensionTask.DimensionForUnit() and then check if the retrieved dimension is not null and is contained in supportedDimensions

[georgeDaskalakis]

@Yuri05 yeap this is a better solution to avoid code duplication - and it would have exactly the same result. If no objections I would implement this.

[PavelBal]

No objections.

[msevestre]

Yes. I thought this was the code used indeed. Let's update

[PavelBal]

This is NOT fixed

UnitCaseInsensitive.xlsx

the unit is "µm", which should be recognized as the dimension "Length". however, this is being recognized as "Concentration (molar)", probably because it thinks it is "µM"

[georgeDaskalakis]

Reopened because currently the fix only works for units in the headers, but the behaviour is still the old faulty one when the unit comes from the content of a column.