Given the update, I have just backed out to sqlite3 (3.34.1) and upgraded 
to weewx 4.9.1 and I can see I'm getting the same result with the 
changes.   Excellent coding Tom.

# python3 -c "import sqlite3; print(sqlite3.sqlite_version)"
3.34.1

obs_lookup is wind
Start time is 1665440100 and end time is 1665442800
x_domain entries are 1665440100 and 1665442800
archive is <weewx.manager.DaySummaryManager object at 0xb2d298f8>
aggregate_type is vecdir
aggregate_interval is 900
([1665440100, 1665441000, 1665441900], 'unix_epoch', 'group_time')
([355.0, 357.7404188153435, 2.124825517227194], 'degree_compass', 
'group_direction')

On Monday, October 24, 2022 at 10:34:49 PM UTC+1 James Taylor wrote:

> No problems
>
> I wish!  My day job is mainly working with Red Hat and its derivatives 
> (Rocky / CentOS)
>
> James
>
> On Monday, October 24, 2022 at 10:08:48 PM UTC+1 [email protected] wrote:
>
>> Thanks for checking this, James!
>>
>> Nice to have a Debian whiz on board!
>>
>> On Mon, Oct 24, 2022 at 1:52 PM James Taylor <
>> [email protected]> wrote:
>>
>>> Yes, I was surprised.  Hopefully Debian 12 ( Bookworm) when it comes 
>>> along in 2023 has a newer version as standard
>>>
>>> I just upgraded to the latest version, which I would only recommend for 
>>> those that are comfortable with this sort of thing.   It didn't actually 
>>> take that long to compile on my Pi4
>>>
>>> mkdir sqlite && cd sqlite
>>> wget https://www.sqlite.org/2022/sqlite-autoconf-3390400.tar.gz
>>> tar xvfz sqlite-autoconf-3390400.tar.gz
>>> cd sqlite-autoconf-3390400
>>> ./configure
>>> make
>>> sudo make install
>>> /usr/local/bin/sqlite3 --version # returns 3.39.4 2022-09-29 
>>>
>>> You also need to replace 
>>> /usr/lib/arm-linux-gnueabihf/libsqlite3.so.0.8.6 with the version compiled 
>>> under sqlite-autoconf-3390400/.libs/, so python3 will start using the newer 
>>> instance (Note if somebody is reading this, it is at their risk and 
>>> understand the consequencies
>>>
>>> # python3 -c "import sqlite3; print(sqlite3.sqlite_version)"
>>> 3.39.4
>>>
>>> However it is now working using the example from my original email
>>>
>>> obs_lookup is wind
>>> Start time is 1665440100 and end time is 1665442800
>>> x_domain entries are 1665440100 and 1665442800
>>> archive is <weewx.manager.DaySummaryManager object at 0xb2c55a60>
>>> aggregate_type is vecdir
>>> aggregate_interval is 900
>>> ([1665440100, 1665441000, 1665441900], 'unix_epoch', 'group_time')
>>> ([355.0, 357.7404188153435, 2.124825517227194], 'degree_compass', 
>>> 'group_direction')
>>>
>>> James
>>> On Monday, October 24, 2022 at 8:39:19 PM UTC+1 [email protected] wrote:
>>>
>>>> That's unfortunate. Bullseye has been out for a while, so I had just 
>>>> assumed that it had 3.35.
>>>>
>>>> I think sqlite maintains their own repository. You could try upgrading 
>>>> from that.
>>>>
>>>> On Mon, Oct 24, 2022 at 11:54 AM James Taylor <
>>>> [email protected]> wrote:
>>>>
>>>>> Cool.
>>>>>
>>>>> Just need to find a way to upgrade sqlite3 on Raspberry PI.    The 
>>>>> packaged version is 3.34.1 on is Raspbian 11 (Bullseye) and this is 
>>>>> wanting 
>>>>> at least 3.35.
>>>>>
>>>>> James
>>>>>
>>>>> On Saturday, October 22, 2022 at 10:26:27 PM UTC+1 [email protected] 
>>>>> wrote:
>>>>>
>>>>>> Fixed with commit 407602c 
>>>>>> <https://github.com/weewx/weewx/commit/407602c03d307945284c2f6544f3f500731699ea>,
>>>>>>  
>>>>>> to appear in V4.9.
>>>>>>
>>>>>> On Tue, Oct 18, 2022 at 4:53 PM Tom Keffer <[email protected]> wrote:
>>>>>>
>>>>>>> You're right that the wind direction plots that come with the 
>>>>>>> Seasons skin are not very useful. They might make sense for short time 
>>>>>>> periods that do not use aggregation, but not for longer periods. 
>>>>>>> Unfortunately, the way vecdir is set up now, aggregation intervals have 
>>>>>>> to 
>>>>>>> be multiples of a day. Hence, your example works, but it won't for 
>>>>>>> plots 
>>>>>>> with shorter aggregation intervals.
>>>>>>>
>>>>>>> I've created Issue #800 <https://github.com/weewx/weewx/issues/800> 
>>>>>>> to track.
>>>>>>>
>>>>>>> On Mon, Oct 17, 2022 at 1:52 PM [email protected] <[email protected]> 
>>>>>>> wrote:
>>>>>>>
>>>>>>>> I'm still trying to wrap my head around wind data in WeeWX... Does 
>>>>>>>> that mean that in the Seasons skin, the weekwinddir, monthwindir, and 
>>>>>>>> yearwinddir plots are not a vector average? If yearwinddir plot was 
>>>>>>>> something like below, would it generate a vector average?
>>>>>>>>         [[[yearwinddir]]]
>>>>>>>>             yscale = 0.0, 360.0, 45.0
>>>>>>>>             line_type = None
>>>>>>>>             marker_type = box
>>>>>>>>             marker_size = 2
>>>>>>>>             #[[[[windDir]]]]
>>>>>>>>             [[[[wind]]]]
>>>>>>>>                 aggregate_type = vecdir
>>>>>>>> When I changed the yearwinddir plot to the above, it ran and I got 
>>>>>>>> what looked like a different plot.
>>>>>>>> Thanks. rich
>>>>>>>>
>>>>>>>> On Monday, 17 October 2022 at 09:05:37 UTC-4 [email protected] 
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> Calculating the vector averaged direction requires a vector, so 
>>>>>>>>> the observation type is 'wind', which is a vector, not 'windDir'. The 
>>>>>>>>> aggregation that returns direction from a vector is 'vecdir', so, you 
>>>>>>>>> want:
>>>>>>>>>
>>>>>>>>> (start_ts, stop_ts, dirs) = weewx.xtypes.get_series('wind', 
>>>>>>>>> x_domain, db_lookup(data_binding=binding), 'vecdir', 
>>>>>>>>> aggregate_interval)
>>>>>>>>>
>>>>>>>>> Unfortunately, 'wind' appears only in the daily summaries. This 
>>>>>>>>> means aggregate_interval must be multiples of one day. This 
>>>>>>>>> restriction 
>>>>>>>>> could be relaxed should someone want to write the necessary xtypes 
>>>>>>>>> extension. 
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Sun, Oct 16, 2022 at 6:35 AM James Taylor <
>>>>>>>>> [email protected]> wrote:
>>>>>>>>>
>>>>>>>>>> Hello
>>>>>>>>>>
>>>>>>>>>> Following on from https://github.com/weewx/weewx/issues/798, 
>>>>>>>>>> I've been trying to replicate and understand weewx behaviour when it 
>>>>>>>>>> comes 
>>>>>>>>>> to using get_series and get_aggregate.
>>>>>>>>>>
>>>>>>>>>> So if you are trying to aggregate winddir for graphing purposes, 
>>>>>>>>>> you want to get a wind vector which I understand
>>>>>>>>>>
>>>>>>>>>> Here is my example where my data is being archived every 300 
>>>>>>>>>> seconds
>>>>>>>>>>
>>>>>>>>>> (time_start_vt, time_stop_vt, obs_vt) = weewx.xtypes.get_series( 
>>>>>>>>>> obs_lookup, x_domain, db_lookup(data_binding=binding), 
>>>>>>>>>> aggregate_type, 
>>>>>>>>>> aggregate_interval)
>>>>>>>>>>
>>>>>>>>>> obs_lookup is windDir
>>>>>>>>>> Start time is 1665440100 and end time is 1665442800
>>>>>>>>>> x_domain entries are 1665440100 and 1665442800
>>>>>>>>>> aggregate_type is avg 
>>>>>>>>>> aggregate_interval is 300 
>>>>>>>>>> ([355.0, 355.0, 355.0, 355.480837630687, 0.0, 0.0, 0.0, 0.0, 
>>>>>>>>>> 4.249651034454402], 'degree_compass', 'group_direction') 
>>>>>>>>>> aggregate_interval is 900 
>>>>>>>>>> ([355.0, 118.49361254356234, 1.416550344818134], 
>>>>>>>>>> 'degree_compass', 'group_direction')
>>>>>>>>>>
>>>>>>>>>> I can see for the second group it is returning an average rather 
>>>>>>>>>> than a vector value of around 358.5, but trying to understand how we 
>>>>>>>>>> should 
>>>>>>>>>> be coding it.
>>>>>>>>>>
>>>>>>>>>> If I can change to aggregate of vecdir or vecavg I get a 
>>>>>>>>>> UnknownType error
>>>>>>>>>>
>>>>>>>>>> Any help will be appreciated here.
>>>>>>>>>>
>>>>>>>>>> James
>>>>>>>>>>
>>>>>>>>>> -- 
>>>>>>>>>> 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].
>>>>>>>>>> To view this discussion on the web visit 
>>>>>>>>>> https://groups.google.com/d/msgid/weewx-user/2ed7f92d-c030-45df-a241-2abde9c647een%40googlegroups.com
>>>>>>>>>>  
>>>>>>>>>> <https://groups.google.com/d/msgid/weewx-user/2ed7f92d-c030-45df-a241-2abde9c647een%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>>>>>>> .
>>>>>>>>>>
>>>>>>>>> -- 
>>>>>>>> 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].
>>>>>>>> To view this discussion on the web visit 
>>>>>>>> https://groups.google.com/d/msgid/weewx-user/febe0e12-adee-4e34-b7d9-9f1791d80531n%40googlegroups.com
>>>>>>>>  
>>>>>>>> <https://groups.google.com/d/msgid/weewx-user/febe0e12-adee-4e34-b7d9-9f1791d80531n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>> -- 
>>>>> 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].
>>>>>
>>>> To view this discussion on the web visit 
>>>>> https://groups.google.com/d/msgid/weewx-user/c851366a-f03e-4f41-af4e-688c555df2d0n%40googlegroups.com
>>>>>  
>>>>> <https://groups.google.com/d/msgid/weewx-user/c851366a-f03e-4f41-af4e-688c555df2d0n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>> -- 
>>> 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].
>>>
>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/weewx-user/216b0605-e9e6-4b4a-b87e-6d91cc640d2dn%40googlegroups.com
>>>  
>>> <https://groups.google.com/d/msgid/weewx-user/216b0605-e9e6-4b4a-b87e-6d91cc640d2dn%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>>

-- 
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].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/weewx-user/90ad1dcc-eb9a-45da-a322-99cd32ce7f5en%40googlegroups.com.

Reply via email to