Well, so much for that theory. About 7:35p it crashed again. But unplugging the device that generates that data has allowed it to run normally again. However, based on your response the last record in this output would be the culprit then correct? If my count is correct, both sum and wsum would be considered None when the accumulator is initialized. Any idea what could cause this behavior?
sqlite> select * from archive_day_extraTemp3 where datetime >= 1510640000; dateTime|min|mintime|max|maxtime|sum|count|wsum|sumtime 1510642800|70.52|1510728900|78.26|1510721100|2935.64|40|880692.0|12000 1510729200|65.48|1510746000|78.08|1510779600|16512.56|232|4953768.0|69600 1510815600|67.28|1510842900|71.06|1510842000|2934.08|43|880224.0|12900 1510902000|||||0.0|0|0.0|0 1510988400|74.48|1511058300|83.12|1511039400||45||13500 sqlite> sqlite> select dateTime, inTemp, extraTemp3 from archive where datetime >= 1511057400; dateTime|inTemp|extraTemp3 1511057400|70.6000000000002|75.02 1511057700|70.3999999999998|74.66 1511058000|70.3999999999998|74.66 1511058300|70.2767123287672|74.48 1511058600|70.2| 1511069400|69.8092307692308| A f thought: In my driver if the correct data is not supplied, None is sent to event.record, is this correct, or should it simply not be called in that circumstance? Snippit from my current driver: for sensor in self.sensors: if sensor != 'mac': # print "DB Entry: " + self.sensors[sensor] try: value = esp_records[sensor] except KeyError: value = None # print "Value: " + str(value) event.record[self.sensors[sensor]] = value I never modify the database without going through weewx, unless I am purging old data, so it shouldn't be an issue, but you never know. Thanks Gary On Saturday, November 18, 2017 at 8:39:03 PM UTC-7, gjr80 wrote: > > Yep, that could explain the symptoms. The accumulator for a given day and > for a given observation is at times initialised from that particular day > from that particular observation's daily summary table. If the sum field of > the day and daily summary concerned was None instead of a numeric then the > sum field of that accumulator would be initialised with None instead of 0.0 > and you would see the symptoms you experienced. Rebuilding the daily > summaries would fix such a problem. It will be interesting to see if the > same issue comes back, certainly if you were making changes to the database > schema there is potential for a hiccup to occur, if you are relying on > weeWX to handle all database interraction then such a hiccup should never > occur. > > 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 weewx-user+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.