This year records issue

[freddie]

Today I have had to edit out a "spike" in average wind speed caused by restarting MX on Saturday. As it was a windy day, the spike registered as a monthly record and a record for this year. That was easy to do using the various editors.

While I was looking at the "this year" records I noticed a couple of other discrepancies in the "logfile" column of the page - one for high hourly rainfall on 1st January (42.8 mm recorded against 13:16; daily rainfall total was 3.8 mm) and one for highest wind run on 12th January (6547.8 nautical miles recorded against this date; actual recorded wind run was 465.5 nautical miles; 6547.8 nm wind run would indicate a 24-hour mean speed of about 273 knots ). See image below:

this-year-records.png

Looking at the monthly log file for these particular dates, the data is fine.

What can I do to correct these obviously erroneous figures? I fear they are not stored anywhere - I think they are calculated on the fly when you hit the "Fetch Monthly Log Data" button.

Actually - looking at some of the other "logfile" values I can see some more discrepancies: low "feels like", low windchill, low maximum temperature (a value from December), low daily temp range, high humidity (value is okay, but the date is from December), and high 24-hour rain (not shown on the above image). Errors are reproduced consistently. Perhaps something to do with me using an 09:00 rollover?

EDIT: I've noticed that the erroneous wind run has a date from December 2022 as well.

[mcrossley]

Nothing for you to correct, I suspect like you it is a problem with the 9am rollover. I'll check it out. Now next on my list....

[mcrossley]

Could you zip up your dayfile, and the 2022-23 monthly log files for me please?

[freddie]

Sure. I assume you mean just the Dec and Jan logs?
Thanks for taking a look.

[mcrossley]

Thanks.

Okay, there are some oddities in your Dec 22 log file

Lines 214 & 229 are dated incorrectly as 01/01/23, the time is correct, but some of the data is not.

Those out of sequence dates really mess up the records code. If I remove those lines, then the yearly records from the log files seem sensible.

Now the question is how did they get there - I fear you are going to say you used the MX internal editor to fix some data? Which means that the log editor code may be the root cause.

[freddie]

Thanks Mark.

Yes, it would've been via the editor as I don't do hand edits on the files. I can't remember any specifics unfortunately.

What I do remember, though, is some odd behaviour in the editor that I was going to raise with you next - where records are missing from the editor but are present in the file. I don't recall details, but I'm pretty sure it was for a day in January 2023. I will take a look later today.

I will remove those erroneous lines and see how the records pages perform.

[mcrossley]

Ok. I'll take a look at the records editor using 9am rollover. It will be a few days though, off boating in Wales again tonight.

[freddie]

Removing the erroneous lines fixed the issues in the "this year" records. Thanks for pointing those out.