NA seg_id's in observations csv's

[jsadler2]

After some really weird behavior in joining temp and flow observations I discovered that some of the flow observations in obs_flow_full.csv have nan for both seg_id_nat and subseg_id.

[jsadler2]

this is the file last uploaded in September though, so maybe things have changed.

[jsadler2]

This may be related, maybe not. There are also a handful of rows in obs_temp_full.csv that have nan as the seg_id_nat but have a subseg_id:

[jsadler2]

Okay. I was looking in an outdated spot in Google Drive. I was looking in the UMN-USGS Data Sharing folder. If I instead look at delaware-model-prep/9_collaborator_data/umn/obs_flow_full.csv, that one has no nans

[jsadler2]

So I guess it got fixed.

[jsadler2]

I do still see the same nans in delaware-model-prep/9_collaborator_data/umn/obs_temp_full.csv as above.

[aappling-usgs]

The second screen shot (#32 (comment)) shows behavior that i think was expected at least at one point: subseg 3_1 doesn't share an outlet with any PRMS reach because seg 3 was subdivided into 3_1 (upstream) and 3_2 (downstream). so that's why seg_id_nat is NA for 3_1.

I have no explanation for the first screenshot.

I remember that you and I talked about the second case and fixed the distance matrix to only include those subsegs that match up to seg_id_nat outlets. But I don't think I or anybody else did anything to subset other data files in the same way. Should we?