Gary,

Excellent. Thanks for sussing this out! I'll look forward to Pat's next 
release.

Phil

On Tuesday, September 25, 2018 at 10:04:07 AM UTC-4, gjr80 wrote:
>
> OK, I think I have handle on what is happening here. The cause is as Phil 
> has noted, the last timestamp in each data series in month.json is 
> midnight at the end of the current local day. A commit by Pat on 4 
> September changed that timestamp from being the last timestamp in the 
> archive to being the midnight timestamp. Looking at the releases that Pat 
> has made I think 0.7 will exhibit this behaviour but 0.6 does not. I will 
> drop Pat a line with the details.
>
> Gary
>
> On Tuesday, 25 September 2018 22:14:20 UTC+10, Philip Kutzenco wrote:
>>
>> So, looking at the data in the json file, the last epoch date in the 
>> series is 1537934400000 and translates to September 26 at 4:00 AM UTC. So 
>> that is midnight on September 26 EDT (my time zone). So it sounds like the 
>> aggregated temperature data displayed is for the period prior to that 
>> date/time. But it definitely makes the plot seem inaccurate, especially 
>> when you look at past dates. You'd expect that the aggregated data point to 
>> correspond to the data gathered from the start of the date to midnight just 
>> before the next date starts. I hope that is clear (and that I'm not 
>> confused). 
>>
>> Phil
>>
>>
>>
>> On Tuesday, September 25, 2018 at 7:59:03 AM UTC-4, Philip Kutzenco wrote:
>>>
>>> Gary,
>>>
>>> I think this is the file you want me to post. It is timestamped 
>>> September 25, 2108 at 7:46 AM. I am in the US Eastern Time Zone currently 
>>> observing Daylight Saving Time. So my time zone is EDT. I checked the date 
>>> and time with the date command on my Raspberry Pi running running weewx and 
>>> it is correct.
>>>
>>> Let me know if you wanted a different file.
>>>
>>> Thanks. I appreciate the help.
>>>
>>> Phil
>>>
>>> On Monday, September 24, 2018 at 11:33:58 PM UTC-4, gjr80 wrote:
>>>>
>>>> I suspect the issue will be related to how highcharts is interpreting 
>>>> timestamps in the plot data rather then the wrong data being in the plot 
>>>> data files. I have seen something like this before but it was related to 
>>>> the timezone of the client computer viewing the plots, would seem that is 
>>>> not the issue here though. I am rather pushed for time at the moment but 
>>>> will have a look when I can, though it's been a while though since I have 
>>>> delved into highcharts.
>>>>
>>>> Phil, can you post one of the generated  highcharts data files that is 
>>>> displaying the issue. Would help to know your timezone and at  what time 
>>>> the file was generated as well.
>>>>
>>>> Gary
>>>>
>>>>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to