Implemented in commit ea9fe13
<https://github.com/weewx/weewx/commit/ea9fe134a242ddd0ea1df9e1c4aa4ddea1542ff6>

-tk

On Sat, Sep 15, 2018 at 5:38 AM Thomas Keffer <[email protected]> wrote:

> Thanks, Chris
>
> I worry that this approach doesn't cover the case where the "previous"
> record has, say, a timestamp far in the future. This patch would cause all
> of the remaining records to get thrown out, but the "future" record to be
> kept.
>
> In any case, it's way better than what we have, which allows negative
> intervals.
>
> -tk
>
>
>
> On Fri, Sep 14, 2018 at 3:53 PM Chris Manton <[email protected]> wrote:
>
>> I suppose if the archived packets are presented out of order there could
>> be a negative interval...
>>
>> I have not seen this but the wmr200 is full of wonderful surprises.
>>
>> One potential drive-by fix is to drop all packets satisfying this failing
>> condition...here is one untested fix:
>>
>>
>> $ git diff
>> diff --git a/bin/weewx/drivers/wmr200.py b/bin/weewx/drivers/wmr200.py
>> index eaab5a14..62b743e0 100644
>> --- a/bin/weewx/drivers/wmr200.py
>> +++ b/bin/weewx/drivers/wmr200.py
>> @@ -1943,7 +1943,13 @@ class WMR200(weewx.drivers.AbstractDevice):
>>                  timestamp_packet_interval = timestamp_packet_current \
>>                          - timestamp_packet_previous
>>
>> -                if pkt.timestamp_record() > (timestamp_packet_previous
>> +                if timestamp_packet_interval < 0:
>> +                    loginf(('genStartup() Discarding received archive'
>> +                        ' record that presented out-of-order'
>> +                        ' current_threshold:%s previous_timestamp:%s')
>> +                        %
>> (weeutil.weeutil.timestamp_to_string(pkt.timestamp_record())
>> +
>>  weeutil.weeutil.timestamp_to_string(timestamp_packet_previous)))
>> +                elif pkt.timestamp_record() > (timestamp_packet_previous
>>                                               + self._archive_threshold):
>>                      loginf(('genStartup() Discarding received archive'
>>                              ' record exceeding archive interval cnt:%d'
>>
>> On Friday, September 14, 2018 at 6:07:48 AM UTC-7, Thomas Keffer wrote:
>>>
>>> Good.
>>>
>>> Looking through the WMR200 driver code (I didn't write it), it appears
>>> that a positive value for 'interval' is not guaranteed. I'll ask the author
>>> to take a look at it.
>>>
>>> -tk
>>>
>>> On Fri, Sep 14, 2018 at 5:44 AM Rahul Bedi <[email protected]> wrote:
>>>
>>>> Yup found a record with -ve interval, changed that to 5 as thats what I
>>>> am using.
>>>>
>>>> And problem fixed !!
>>>>
>>>> Thanks a lot.
>>>>
>>>> Not sure how this bad data entered into the db.
>>>>
>>>> Will be great if there is utility that finds such errors in the db and
>>>> flags them.
>>>>
>>>> - Rahul
>>>>
>>>> On Fri, Sep 14, 2018 at 5:48 PM Thomas Keffer <[email protected]> wrote:
>>>>
>>>>> I think the thing to do is to figure out how a negative value got in
>>>>> there, fix that, then rebuild the daily summaries.
>>>>>
>>>>> One possibility is that a record in your database has a negative value
>>>>> for 'interval' --- the time length of an archive record. Let's check that.
>>>>>
>>>>> *sqlite3 /var/lib/weewx/weewx.sdb*
>>>>> sqlite> *SELECT datetime(dateTime, 'unixepoch','localtime'), interval
>>>>> FROM archive WHERE interval<0;*
>>>>>
>>>>> -tk
>>>>>
>>>>> On Fri, Sep 14, 2018 at 5:06 AM Rahul Bedi <[email protected]> wrote:
>>>>>
>>>>>> Here is what I get:
>>>>>>
>>>>>> 2018-08-02 00:00:00|2851473.0|86400
>>>>>> 2018-08-03 00:00:00|2139447.0|86400
>>>>>> 2018-08-04 00:00:00|2867211.0|86400
>>>>>> 2018-08-05 00:00:00|3306195.0|86400
>>>>>> 2018-08-06 00:00:00|3167394.0|86400
>>>>>> 2018-08-07 00:00:00|2924283.0|86400
>>>>>> 2018-08-08 00:00:00|2255208.0|86400
>>>>>> 2018-08-09 00:00:00|2797065.0|86400
>>>>>> 2018-08-10 00:00:00|2693187.0|86400
>>>>>> 2018-08-11 00:00:00|2981601.0|86400
>>>>>> 2018-08-12 00:00:00|3869136.0|86400
>>>>>> 2018-08-13 00:00:00|3632787.0|86400
>>>>>> 2018-08-14 00:00:00|2651829.0|86400
>>>>>> 2018-08-15 00:00:00|2104410.0|86400
>>>>>> 2018-08-16 00:00:00|3682674.0|86400
>>>>>> 2018-08-17 00:00:00|3499992.0|86400
>>>>>> 2018-08-18 00:00:00|3855531.0|86400
>>>>>> 2018-08-19 00:00:00|3875121.0|86400
>>>>>> 2018-08-20 00:00:00|3915861.0|86400
>>>>>> 2018-08-21 00:00:00|3523719.0|86400
>>>>>> 2018-08-22 00:00:00|1637909.95179334|85620
>>>>>> 2018-08-23 00:00:00|506702.215048778|25380
>>>>>> 2018-08-24 00:00:00|523285.18735593|101340
>>>>>> 2018-08-25 00:00:00|2047231.90482286|92280
>>>>>> 2018-08-26 00:00:00|2207680.59885907|85800
>>>>>> 2018-08-27 00:00:00|2112653.43729687|78000
>>>>>> 2018-08-28 00:00:00|1927695.38113441|85200
>>>>>> 2018-08-29 00:00:00|1696867.80614767|76740
>>>>>> 2018-08-30 00:00:00|1892572.6564943|85500
>>>>>> 2018-08-31 00:00:00|1669207.56530998|86100
>>>>>> 2018-09-01 00:00:00|1882876.90336366|85080
>>>>>> 2018-09-02 00:00:00|1552040.97205686|86100
>>>>>> 2018-09-03 00:00:00|1550916.37957731|86100
>>>>>> 2018-09-04 00:00:00|1093704.34743737|85740
>>>>>> 2018-09-05 00:00:00|1688380.4274756|86100
>>>>>> 2018-09-06 00:00:00|2349485.94738889|86100
>>>>>> 2018-09-07 00:00:00|2676307.36370007|85560
>>>>>> 2018-09-08 00:00:00|2198034.86178985|86100
>>>>>> 2018-09-09 00:00:00|-7107216.29856356|-302280
>>>>>> 2018-09-10 00:00:00|288713.672694377|85560
>>>>>> 2018-09-11 00:00:00|351094.880509644|85860
>>>>>> 2018-09-12 00:00:00|441249.222795585|85500
>>>>>> 2018-09-13 00:00:00|60033.1634844758|85620
>>>>>> 2018-09-14 00:00:00|21222.2555021576|62820
>>>>>>
>>>>>> I am guessing that the negative values above may be the culprit. How
>>>>>> are these populated in the archive_day_wind table or should I just edit
>>>>>> these values out ?
>>>>>>
>>>>>> - Rahul
>>>>>>
>>>>>>
>>>>>> On Fri, Sep 14, 2018 at 5:19 PM Thomas Keffer <[email protected]>
>>>>>> wrote:
>>>>>>
>>>>>>> You're on the right track: there must be an bad value in your
>>>>>>> database. Did you, by any chance, use a visual sql editor? They often 
>>>>>>> put
>>>>>>> null strings in the database.
>>>>>>>
>>>>>>> In any case, the field 'wsquaresum' and 'sumtime' are used by rms.
>>>>>>> So, try this:
>>>>>>>
>>>>>>> *sqlite3 /var/lib/weewx/weewx.sdb*
>>>>>>> sqlite> *SELECT datetime(dateTime, 'unixepoch','localtime'),
>>>>>>> wsquaresum, sumtime FROM archive_day_wind WHERE dateTime > 1533081600;*
>>>>>>>
>>>>>>> This will show all values from (roughly) 1-August (about 45 values).
>>>>>>> If the bad value is before that, we'll have to use a different strategy,
>>>>>>> rather than go looking for a needle in a haystack.
>>>>>>>
>>>>>>> -tk
>>>>>>>
>>>>>>> On Thu, Sep 13, 2018 at 11:35 PM Rahul Bedi <[email protected]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> I found rms used in the statictics.ico and hilo.ico files in the
>>>>>>>> Seasons directory. removed those segments. Diff as follows:
>>>>>>>>
>>>>>>>> $ diff statistics.14092018.inc statistics.inc
>>>>>>>>
>>>>>>>> 134,140d133
>>>>>>>> <         <tr>
>>>>>>>> <           <td class="label">$obs.label.wind_rms</td>
>>>>>>>> <           <td class="units">$unit.label.wind</td>
>>>>>>>> <           #for $archive in $archive_data
>>>>>>>> <           <td class="data
>>>>>>>> new_row">$archive.wind.rms.formatted</td>
>>>>>>>> <           #end for
>>>>>>>> <         </tr>
>>>>>>>>
>>>>>>>> $ diff hilo.inc hilo.14092018.inc
>>>>>>>>
>>>>>>>> 132a133,140
>>>>>>>> >       <tr>
>>>>>>>> >         <td class="label">$obs.label.wind_rms</td>
>>>>>>>> >         #for $archive in $archive_data
>>>>>>>> >         <td class="data new_row hilo_$archive[0]">
>>>>>>>> >           $archive[1].wind.rms.formatted</td>
>>>>>>>> >         #end for
>>>>>>>> >         <td class="units">$unit.label.wind</td>
>>>>>>>> >       </tr>
>>>>>>>>
>>>>>>>> Apart from this any file referring to rms were exactly the same as
>>>>>>>> the original install. Attaching skin.conf with this message.
>>>>>>>>
>>>>>>>> As expected the problem is gone but rms is not getting reported.
>>>>>>>> Which means the issue still stays.
>>>>>>>>
>>>>>>>> Could you help me with what fields in the archive are referred to
>>>>>>>> by rms ? I can try to manually check for any value discrepancy.
>>>>>>>>
>>>>>>>> Thanks.
>>>>>>>>
>>>>>>>> - Rahul
>>>>>>>>
>>>>>>>>
>>>>>>>> On Fri, Sep 14, 2018 at 9:14 AM Rahul Bedi <[email protected]>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> Thanks Gary.
>>>>>>>>>
>>>>>>>>> Charges were done about a month ago but have seen no errors till
>>>>>>>>> this episode.  I will do a diff with the original skin files and post 
>>>>>>>>> it
>>>>>>>>> here.
>>>>>>>>>
>>>>>>>>> Thanks.
>>>>>>>>>
>>>>>>>>> - Rahul
>>>>>>>>>
>>>>>>>>> On Fri 14 Sep, 2018, 7:35 AM gjr80, <[email protected]> wrote:
>>>>>>>>>
>>>>>>>>>> On Friday, 14 September 2018 11:30:42 UTC+10, Rahul Bedi wrote:
>>>>>>>>>>>
>>>>>>>>>>> As I have made no code or config changes since the day before, I
>>>>>>>>>>> believe it doing this for the right data.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> This sounds to me like you have changed some of the seasons skin
>>>>>>>>>> files. The error is coming form the use of a .rms tag in one of
>>>>>>>>>> the seasons .tmpl or .inc files. The seasons skin only uses the
>>>>>>>>>> .rms tag with the WeeWX wind field. As Tom indicated .rms only
>>>>>>>>>> works against the wind field, if your changes have used the .rms
>>>>>>>>>> tag against another field then this could be the cause. I suggest 
>>>>>>>>>> you go
>>>>>>>>>> through all of your changes to the seasons skin files and see what 
>>>>>>>>>> fields
>>>>>>>>>> are being used with the .rms tag. Until we know for sure, and
>>>>>>>>>> given you seem to have made changes to the skin, we will likely be 
>>>>>>>>>> chasing
>>>>>>>>>> our tails.
>>>>>>>>>>
>>>>>>>>>> Gary
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> You received this message because you are subscribed to a topic
>>>>>>>>>> in the Google Groups "weewx-user" group.
>>>>>>>>>> To unsubscribe from this topic, visit
>>>>>>>>>> https://groups.google.com/d/topic/weewx-user/LWmSpcVfdmo/unsubscribe
>>>>>>>>>> .
>>>>>>>>>> To unsubscribe from this group and all its topics, send an email
>>>>>>>>>> to [email protected].
>>>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>>>
>>>>>>>>> --
>>>>>>>> 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.
>>>>>>>>
>>>>>>> --
>>>>>>> You received this message because you are subscribed to a topic in
>>>>>>> the Google Groups "weewx-user" group.
>>>>>>> To unsubscribe from this topic, visit
>>>>>>> https://groups.google.com/d/topic/weewx-user/LWmSpcVfdmo/unsubscribe
>>>>>>> .
>>>>>>> To unsubscribe from this group and all its topics, send an email to
>>>>>>> [email protected].
>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>
>>>>>> --
>>>>>> 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.
>>>>>>
>>>>> --
>>>>> You received this message because you are subscribed to a topic in the
>>>>> Google Groups "weewx-user" group.
>>>>> To unsubscribe from this topic, visit
>>>>> https://groups.google.com/d/topic/weewx-user/LWmSpcVfdmo/unsubscribe.
>>>>> To unsubscribe from this group and all its topics, send an email to
>>>>> [email protected].
>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>
>>>> --
>>>> 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.
>>>>
>>> --
>> 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.
>>
>

-- 
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