-901 values not null in 2.0

[BrittCorley]

1.4.3 fills null cells with -901 when exporting, so most text files used in 1.4.3 are creating false values of -901 in 2.0 that should be blank. This is most common in the occupancy imports. The last row of the occupancy has data for car and other values that are commonly filled with -901. This can be seen in the workshop data.

[rnugent3]

this feels like it could cause issues in results if we're treating -901 as an actual structure or occ type attribute value and attempt to calculate damage based upon -901 for something

[rnugent3]

for example - if an OSVR is blank but there is an OSVR uncertainty parameter, we'll try to use -901 for the central tendency, as mentioned in #1115

[rnugent3]

@BrittCorley I need study data to replicate this issue.

Generally, the importer project handles the -901 correctly. I confirmed with the workshop occ type data. In the screen shot, you can see that there are -901s for vehicle and other, and in the occ type in 2.0, the tabs for vehicle and other are not selected.

I need to know when the -901s show up in an occ type in 2.0 and when 2.0 tries to use -901 in the compute. In the example you provided (workshop data), I don't see the -901s making it into the compute.

[rnugent3]

I removed the priority label until it can be shown that the software is behaving badly