getOxygenSaturationHistory on a physical device

I've been trying to make SensorHistory.getOxygenSaturationHistory work on a physical device, and the values in that iterator are completely absurd, even with all the check and filters, besides the bad values, the results are simply unpredictable. Consecutive calls with the same params return different results.

I've read the bug reports and several discussions, so I'm aware of the issues, but I'm curious what exactly does Garmin use in their preinstalled PulseOx widget?

Top Replies

jim_m_58 over 4 years ago in reply to Joe Berger +1 verified

You can see the last reading in Sensor, but as noted else where SensorHistory only gets a new reading about once an hour. What you see there, at least on a 945 is 192 samples - one sample an hour for 8…

All Replies

0 jim_m_58 over 4 years ago

There are byg reports about this. Last I checked on a real device, the readings were always oldest first, even if you specify newest first, Seems sothing isn't working right with seraching the bug reports, but it was reported sometime around last fall,

0 Joe Berger over 4 years ago in reply to jim_m_58

Thanks, Jim. yeah, I've read the reports, however, it's not just an issue of order - that would be fairly simple to deal with. The issue is - that iterator returns random garbage, that's different with every call. Some values are NaN, some are ridiculously huge integers, some are negative floats or doubles.

0 jim_m_58 over 4 years ago in reply to Joe Berger

There are also bug reports of bad values, but that could vary based on the device and FW. The FW has this data, but in CIQ there's things needed like turning into mc objects..

0 Joe Berger over 4 years ago in reply to jim_m_58

Got it. Thanks. Search functionality on this forum has been broken for a few days, so difficult to search...

I think I figured out where to get the correct value. Activity has currentOxygenSaturation, and the value seems to persist nicely. I think I'm just going to build my own iterator by saving those values in an array and iterate over it to plot the graph.

+1 jim_m_58 over 4 years ago in reply to Joe Berger

You can see the last reading in Sensor, but as noted else where SensorHistory only gets a new reading about once an hour. What you see there, at least on a 945 is 192 samples - one sample an hour for 8 days, which is what you see in the widget.

0 jim_m_58 over 4 years ago in reply to jim_m_58

Been running my old test app on a 945 with the latest FW for a few hours, and still the same. History has 192 samples, and the new samples are added at the end, and the oldest drop off the front.

So right now, the 192,191,190,and 189th samples have data.

0 Joe Berger over 4 years ago in reply to jim_m_58

I believe you. I was running for a day yesterday on my vivoactive 4, and the iterator would return random values with each call, most of them NaN. I saw someone report the same fundings.

Must be hardware-dependent...

Anyway, I implemented my own iterator. Seems to be working fine.

Thank you for all your help!

0 jim_m_58 over 4 years ago in reply to Joe Berger

NaN? As in "not a number"? I see nulls for the first "x" samples", so the same thing.

0 Joe Berger over 4 years ago in reply to jim_m_58

Yep, not a number. I was checking for nulls before operating on values, and from time to time app would crash with this error. The reason - huge numbers with too many digits before and after decimal point. Someone else posted here with the same issue.