New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Daily summary for 'wind' has wrong values for dirsumtime
#642
Comments
So if I understand this correctly the fix is simply to add That would fix any upgrades from 4.2 to the next release but what solution do we want to do for those on 4.3? Do we need to have daily summaries v4 so that we can reweight from the (erroneous) v3 to v4? |
That does seem to be the problem. I guess we will need a V4. Sigh. Thanks so much. I was so steamed about the problem I was having a hard time seeing anything. |
OK. I’ve created branch dirsumtime. Done the easy bit, will work on the V4 daily summaries bit today. |
Been a while since I worked with patching daily summaries to v2.0 so thinking aloud to make sure v4.0 will cover all possibilities. Let's assume v4.0 daily summaries is release as WeeWX v4.4.0.
So I think if we just rebuild the >= 1 June 2020 daily summaries on encountering v3.0 daily summaries we should be write. The other cases will be taken care of with the existing code. One question springs to mind; how 'good' do we want to be when rebuilding the daily summaries, do we want to preserve highs and lows (will require more code) or not (simple code, just a call to |
OK, so the 3:30am start this morning clouded my brain. If we encounter v3.0 daily summaries the fix is not to rebuild the daily summaries but rather re-weight again from 1 June 2020 but this time with the updated |
Not my day, I've committed changes to branch My apologies for the intermingled 40 odd changes to I've yet to test the upgrades from various database versions, will get onto that next. |
Having messed this up in the first place, I'm in no position to offer advice, but my instinct says to keep the V2 to V3 machinery, but this time get it right. Then invoke it for databases that are at either V2 or V3, ending up with a database that is unambiguously at V4. That's slower than a specialized V3 to V4 update, but probably more bulletproof, with one less thing that we have to test. Final step is to add a regression test to |
I don't think you really mean that.
That is what I have done, the old
Started down that path but was just going to duplicate too much code. Wasn't too worried about time, your method for updating from v2.0 to v3.0 was a big improvement over the method used patching V1.0 to v2.0.
Oh joy. |
Limited testing complete, I'm satisfied it's doing as it should. Method and results
Made a few assumptions in testing:
Will move onto the regression test and documentation. |
Before I go down the wrong track re regression testing can I confirm your intent. Current regression testing just checks |
It's a regression test, so it's OK to hard wire in the correct values for |
Merged into Thanks so much for taking care of this, Gary. It was driving me crazy. |
Using V4.2, we get these values for
archive_day_wind
:After upgrading to V4.3 and running the reweighting patch, we get these values:
The values for
xsum
andysum
got updated, but notdirsumtime
.The text was updated successfully, but these errors were encountered: