Page 1 of 1

v6.23.4

Posted: Tue 16 Jan 2024 10:52 pm
by PaulMy
Hi Hans,
In reading the release I notice
General: The limit of data only after 27 Januari 2004 (start of Cumulus) has been lifted (under the condition that the logfile records have the length of todays records). CUtils can now read whatever number of years (current record is 1 Januari 1979)
and that reminded me of the issues and discussions in the past with my older data and wondered if this release may have addressed that. After doing thew website runs I checked the logs (a lot more information and more concise than last time I looked) and noticed the same errors still showing from the years 2008 to 2011. I've tried to find the error in my log files but didn't. I have attached a txt file where I have inserted some of the errors, and put it in a table comparing to my dayfile for 01/04/09 and also some other dates (I manipulated the data items to show in a neat column for easier comparing).

Not a concern from me, just mentioning following your release comments.

Enjoy,
Paul

Re: v6.23.4

Posted: Wed 17 Jan 2024 3:58 am
by HansR
Thanks for the effort Paul, I'll analyse and come back to this.

I kind of feared there were still issues but the idea is that CUtils now can read very far back. Point is that far back read is done - and has to be done - with datafiles which have the modern format (reworked data). I reworked the error handling so it does not surprise me some old errors may show up but I'll definitely see if I can handle it differently.

Re: v6.23.4

Posted: Wed 17 Jan 2024 8:40 am
by HansR
Hi Paul,

Looking at your files there are two remarks:

1) Your dayfile.txt does not pose any problems most likely because you ran CreateMissing.
2) Your monthly logfiles of 2008 - 2010 don't have solar on which the reading fails. Which is correct and it has always been that way.

Note that the previous CUtils had a barrier built in for any file before 2004, that barrier is what has been lifted. But CUtils can't read data which is not there. The only option would be to - programmatically - add missing fields with a value (most likely 0) in your own local format. Otherwise the errors will remain for each file opened. And correctly so I think.

Re: v6.23.4

Posted: Wed 17 Jan 2024 11:06 am
by HansR
What I just think of is that it is a piece of cake to just fill up the line by CUtils with zero and ignore the errors, just give a message the line is filled and that is it. I assume you can live with that? Possibly even store the file as if it is original after filling it up 8-)