I can see either answer in this case. Agree with Mark about NULL vs. zero. Tom's words in the wiki recommend NULL (link <https://github.com/weewx/weewx/wiki/Cleaning-up-old-bad-data>).
On Friday, October 31, 2025 at 3:31:16 AM UTC-7 Mark Jenks wrote: > That is exactly what I've done in the past. Just find the bad data and > NULL it out. NULL says no data, 0 says no rain. There is a difference. > > There is no good reason to edit it to try and figure out what it was, > unless there was some huge event that you failed to capture accurately. > > On Thursday, October 30, 2025 at 12:54:31 PM UTC-5 Jeff A. D. wrote: > >> For the sake of accuracy, I think the quickest and easiest way would be >> to just go through and select all the dates with questionable data in the >> database and set all the rain and rain rate data to null, rather than zero, >> and then rebuild dailies. That should tell you you have no data for those >> times, rather than indicating no rain. (It should show "N/A", rather than >> 0, for those dates on the Climatological Summary.) >> >> On Thursday, October 30, 2025 at 9:44:57 AM UTC-6 vince wrote: >> >>> I would just identify which 5-minute archive periods have bad data, then >>> zero out the rain and rainRate fields out for those 5-minute period >>> records. That would be close enough for me. You seem to have something >>> far more complicated in mind, so best of luck. >>> >>> On Thursday, October 30, 2025 at 4:18:31 AM UTC-7 S Phillips wrote: >>> >>>> >>>> So the data which I need to focus on is the rain data that is held in >>>> the archive table and once I can determine the bad values I can then >>>> rebuilt the daily which should correct the issue. Since I live so close >>>> to >>>> where the official readings are kept (~1.5 miles) I can use that data as a >>>> reference. I know that there will be variation but extremes differences >>>> should be easy to spot. For example, here is July 2016 from NOAA and my >>>> PWS >>>> where you can see the extreme variations. >>>> >>>> [image: Combined 2016-07 copy.png] >>>> On Wednesday, October 29, 2025 at 7:52:27 PM UTC-5 vince wrote: >>>> >>>>> Forgot to answer your question - if you rebuilt-daily then your bad >>>>> data is in the archive table (which is used to generate the summary table) >>>>> >>>>> Expecting your local rain total in an extreme event to match anybody >>>>> else is a bad idea. Microclimates can have different answers across the >>>>> street from the other station, let alone from one miles away. >>>>> >>>>> You certainly can fix up the rainRate item in your archive table, or >>>>> at least zero it out, but I would suspect your rain field (rain in that >>>>> usually 5 minute period) likely needs similar cleanup. >>>>> >>>>> -- 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 visit https://groups.google.com/d/msgid/weewx-user/43618a8a-4753-4651-9635-663aef9e1808n%40googlegroups.com.
