Cumulus Support

wurzelmac wrote: ↑Wed 09 Oct 2024 11:06 am CreateMissing.dll is missing in the latest distribution (b4039) - I'll try to copy an instance of b4029 of this file...

Update: Not working...

Update: Got it (CreateMissing.dll) running now...

Is it possible, that there is some issue with the comma place in High hourly rain and High 24 hour rain (1560,3 instead of maybe 156,03 for example)?

No, that is the time the log format changes. See...
Now WC did some "funny stuff" at this point. You can see the new fields Pm, Py contain negative values(?!), and the P field was decremented by the same amount that the Pm/Py went negative. Sum them and you are back to the previous P total of 1342.65

However, that only lasts until 11:55, then the format goes back to the old version again. Was that a test install of WC that did not work properly?

What I'd suggest is that you delete those "bad" entries from 2011/10/25 11:27 to 11:54 from your WC file (lines 27196 - 27224) and run the conversion again.

In that same file, there is again another batch of "weird" lines - 28th 15:12 thru 16:12 (lines 30624-30679) I'd delete those as well.

Hello Mark, thank you for your work and your help. I will follow your advice and try to remove all bad entries that were detected by CMX. Maybe its a mammouth work, but there is a long, cold winter standing in front of the door.
Looks like WeatherCat itself can handle this "funny stuff" because I can't see those values in the records sections of WeatherCat.

However - thanks again, I think you can go to a final release with the converter.

Hello again - another question (maybe) belonging to the converter. Solar radiation seems to be imported from WC data correctly (I added Solar and UV in 2013). But as you can see, sunshinehours are only displayed when I first run CMX, also there are these strange UV indices far above of whats normal (I think all the indices are having a wrong decimal place [56 instead of 5,6 - 80 instead of 8,0 ...]).

This is NOT an urgent thing, just want to tell you.

I'll check the UV-I thing, does WC store the value as X10 or is it my conversion that it the problem I wonder.

I'll check the sunshine hours too.

Hello Mark,
I do have nearly exact the same data set in my log file, nevertheless the values of UV are only correct since starting CMX.

20/05/24,13:10,16.9,56,8.1,10,11,128,0.0,0.0,1009.1,266.2,22.1,49,11,16.9,16.9,8.0,1202,1.68,224.28,14.6,958,0.0,128,0.0,0.0,15.8,17.4

Another screenshot from my test installation of b4039 on another machine (Macbook Air, installed on oct 12th) - and here is the same issue: UV and sunshine hours correct computed only since first run of CMX...

Could you post your August log file please?

Here you are. I noticed that I have two ExtraLog files for 2024_08, one with .txt and another with log.txt, in 09 an 10 there are only those with .txt, no log attached, see screenshot - Thanks in advance!

And thanks for putting all those posts into the correct thread.

Great thanks.

Looking at the UV-I issue.

The monthly log files created by the importer are OK. But the imported values show value * 10 in the day file.

So, it must be the CreateMissing utility that has the problem with UV-I... Investigating... It has (not using invariant culture for solar data), I've just released an update to fix that - version 2.0.2

Are those Extra logs without the "log".txt old ones from early versions of the importer?

Hello, those files with *log.txt* in the file name were created by the importer, the other three (*.txt*) were created by CMX itself.

To the CreateMissing utility: In my case, have I to do all imports from WeatherCat again, run CreateMissing, create new DayFiles, and work over all entries in the EditRecords section within the dashboard?

Edit: And is there some news about the missing sunshine hours?

Thanks in advance,

WeatherCat does not seem to record sunshine hours.

CreateMissing does not calculate them, nor at the moment does the importer. Iirc, your WC log files are recording data every minute - is that standard, or user selectable?

If it is every minute, then I guess the importer could calcuate the sun hours using the settings in cumulus.ini

However if the WC data could be say every 10 minutes, then the calcuations would be less accurate.

Very interesting, maybe WeatherCat calculates the sunshine hours internally. There are special web tags within WeatherCat for this.