Re-generating a curve which has an error / debug

[Invisible Man]

Hi,

There's something strange, and I'd like your help to debug the issue.
The outside temperature curve of my weather station has a cut during the night of Aug 21 and Aug 22. But, to my understanding, the data exists in the SQLite database!
Why?

How can I trigger the curve to be re-generated ?

This is the content of the database. I cut off some of the long output, but you can see I always have an outside temperature. No error.

select datetime(dateTime,'unixepoch', 'localtime'),outTemp from archive where dateTime > 1692640000 and dateTime < 1692680000;
...
2023-08-21 23:00:00|28.1666666666667
2023-08-21 23:05:00|28.1666666666667
2023-08-21 23:10:00|28.25
2023-08-21 23:15:00|28.2222222222222
2023-08-21 23:20:00|28.1111111111111
2023-08-21 23:25:00|28.0
2023-08-21 23:30:00|28.0
2023-08-21 23:35:00|28.0
2023-08-21 23:40:00|28.0
2023-08-21 23:45:00|28.0
2023-08-21 23:50:00|27.8888888888889
2023-08-21 23:55:00|27.8888888888889
2023-08-22 00:00:00|28.1666666666667
2023-08-22 00:05:00|28.3055555555556
2023-08-22 00:10:00|28.1481481481481
2023-08-22 00:15:00|28.1666666666667
2023-08-22 00:20:00|28.4444444444444
2023-08-22 00:25:00|28.1666666666667
2023-08-22 00:30:00|27.8888888888889
2023-08-22 00:35:00|28.1111111111111
2023-08-22 00:40:00|28.25
2023-08-22 00:45:00|28.4259259259259
2023-08-22 00:50:00|28.4444444444444
2023-08-22 00:55:00|28.3888888888889
2023-08-22 01:00:00|28.8055555555556
2023-08-22 01:05:00|28.0740740740741
2023-08-22 01:10:00|27.1666666666667
2023-08-22 01:15:00|27.2777777777778
2023-08-22 01:20:00|27.1111111111111
2023-08-22 01:25:00|27.3055555555556
2023-08-22 01:30:00|27.3703703703704
2023-08-22 01:35:00|26.9444444444444
2023-08-22 01:40:00|26.8518518518519
2023-08-22 01:45:00|26.7777777777778
2023-08-22 01:50:00|26.8518518518519
2023-08-22 01:55:00|26.75
2023-08-22 02:00:00|26.6481481481481
2023-08-22 02:05:00|26.6111111111111
2023-08-22 02:10:00|26.3888888888889
2023-08-22 02:15:00|26.2222222222222
2023-08-22 02:20:00|26.1666666666667
2023-08-22 02:25:00|26.2962962962963
2023-08-22 02:30:00|26.25
...

• Weewx 4.10.2 on a Raspberry Pi
• Actually I have the same problem on outside humidity and barometer.

[michael.k...@gmx.at]

Don't forget to run https://weewx.com/docs/latest/utilities.htm#Action_--rebuild-daily for the particular days.

Weekly images are not generated every archive interval. If the data is there, weekly images should be there by now (14:00 MESZ) If not, check the log for error and/or delete the picture in /var/www/whereeverthereportis to force a new generation of the picture.

[Invisible Man]

Ok, so I removed the weektempdew.png files I found in my weewx directory (to be sure to get new ones), and ran the command.

Unfortunately, there's still a hole in the image (?!) and there are no errors in the logs nor from the output of the command.

weewx@vegan:~$ /usr/bin/wee_database --rebuild-daily --from=2023-08-21 --to=2023-08-22
Using configuration file /etc/weewx/weewx.conf
Using database binding 'wx_binding', which is bound to database 'archive_sqlite'
Daily summaries from 2023-08-21 through 2023-08-22 inclusive will be rebuilt.
Proceed (y/n)? y
Rebuilding daily summaries in database 'weewx.sdb' ...
Processed 576 records to rebuild 2 daily summaries in 1.45 seconds
Rebuild of daily summaries in database 'weewx.sdb' complete

Proof it has been re-generated:

weewx@vegan:~$ ls -al ./public_html/weektempdew.png
-rw-r--r-- 1 root root 14695 Aug 23 14:15 ./public_html/weektempdew.png

[michael.k...@gmx.at]

Forget the rebuild stuff, if you haven't filled back the data. In my post above I assumed you did so, sorry for that if it confused you.

Is the public_html directory the one you are viewing the picture from, or is a data transfer to a remote web space in between?

[gjr80]

The missing data in your plot is from midnight 20 August through until just after sunrise on 21 August not from 22 August as per your database extract. Try dropping each timestamp in your query by 86400.

Gary

[Invisible Man]

Ah ha... you're right... and then things make sense, because indeed I don't have the output temperature for those dates. So, the curve are correctly generated. Now I need to figure out why I had no data at that time, knowing that I had no sensor error warning in the logs...

sqlite> select datetime(dateTime,'unixepoch', 'localtime'),outTemp from archive where dateTime > 1692553600 and dateTime < 1692593600;
2023-08-20 19:50:00|31.75
2023-08-20 19:55:00|31.6481481481481
2023-08-20 20:00:00|31.6111111111111
2023-08-20 20:05:00|31.462962962963
2023-08-20 20:10:00|31.3333333333333

...

2023-08-21 00:00:00|
2023-08-21 00:05:00|
2023-08-21 00:10:00|
2023-08-21 00:15:00|
2023-08-21 00:20:00|
2023-08-21 00:25:00|
2023-08-21 00:30:00|
2023-08-21 00:35:00|
2023-08-21 00:40:00|
2023-08-21 00:45:00|
2023-08-21 00:50:00|
2023-08-21 00:55:00|
2023-08-21 01:00:00|
2023-08-21 01:05:00|
2023-08-21 01:10:00|
2023-08-21 01:15:00|
2023-08-21 01:20:00|
2023-08-21 01:25:00|
2023-08-21 01:30:00|
2023-08-21 01:35:00|