Some profiles start before the previous profile ends

[kwilcox]

I've noticed that in a very small number of Deployments there are profiles that start before the previous profile has ended. This is based on the precise_time(obs) variable and not the time(profile) variable. I haven't identified all of the offending Deployments, but here are a few:

• bass-20150706T151619Z
• Murphy-20150809T1355Z
• Stommel-20150805T0000Z
• unit_308-20150701T222927Z
• unit_540-20150701T181602Z

Here is a notebook showing the problem. The file I'm accessing was downloaded from ERDDAP as a .ncCF - Download a NetCDF-3 CF Discrete Sampling Geometries file (Contiguous Ragged Array).

[lukecampbell]

This is up to the providers. GliderDAC does not have or maintain policies directing continuity for deployments, or overlapping times. We can let them know that it's causing problems for you but as far as I know, we don't have any intention to direct them otherwise.