search

NOAA-PMEL / LAS

Live Access Server
https://ferret.pmel.noaa.gov/LAS/
The Unlicense
13 stars 5 forks source link

SOCAT: selection of cruises for table of cruises used decimated data instead of full data #1711

Open karlmsmith opened 6 years ago

karlmsmith commented 6 years ago

Reported by @karlmsmith on 12 Mar 2015 17:01 UTC Ute reported that selection of N/U/Q-flag cruises in the North Atlantic (region_id = A) in 2012-2014 gave 22 cruises in the table of cruises, but 29 cruises in the correlation viewer.

Investigating this, the 29 cruises includes all the 22 cruises. The six of the seven additional cruises do not have region_id = "A" in the decimated data although there are some in the full data. The seventh cruise 77CN20140531 does have "A" region in the decimated data files, but ERDDAP is not reporting that cruise with the region_id = "A" restriction. So results in the same issue in LAS, although for different reasons.

I would think that LAS should be selecting cruises based on the full data, as is done with the correlation viewer. THe other option would be to restrict decimation to keep any points that change region, but that only fixes this issue for region_id restrictions.

Migrated-From: http://dunkel.pmel.noaa.gov/trac/las/ticket/1704

karlmsmith commented 6 years ago

Modified by @karlmsmith on 12 Mar 2015 17:03 UTC

karlmsmith commented 6 years ago

Modified by @noaaroland on 12 Mar 2015 22:20 UTC

karlmsmith commented 6 years ago

Comment by @karlmsmith on 24 Mar 2015 17:12 UTC So testing with the latest build shows the same number of cruises with the table of cruises, the thumbnail plots, and the correlation viewer when specifying a region_id. But it does not match the number of cruises mentioned in the annotations on the main world plot, which makes sense since the world plot has to use the decimated data.

I think we can just accept this error, since the annotations do say it is using decimated data.

Or we can modify the decimation so changes in region ID keeps the point, but this only will fix this issue for region ID; other restrictions on data values (e.g., season, WOCE flags, etc) will have the same issue and we would go crazy trying to catch them all. But maybe region ID is a special case since QC-ers are divided up by region ID to QC.

karlmsmith commented 6 years ago

Comment by @karlmsmith on 25 Mar 2015 00:16 UTC SocatQCEditor on Xtra updated to latest from the socat branch of baker.