Page 1 of 1

Fixing obviously wrong data

Posted: Wed 19 Jan 2011 9:31 pm
by permaculturegeek
Hi,
we set up a Scientific Sales (Fineoffset rebrand) station on our new land at the start of the month, to gather data on our microclimate before we build. Since we have no power on site yet, we are downloading data weekly.

The latest batch of data has one record which is obviously nonsense - it has a temperature of -25.2C, which is colder than the record low for New Zealand, (it's midsummer!) and also happens to be the negative of our highest temperature.

After backing up the data file, I removed that line of data. However, or records display still shows -25.2 as our lowest temperature, and 1 day of ground frost.

Is there any way of forcing cumulus to recalculate the records file from the data, or do I need to search for the minimum temperature and manually edit the records files?

Re: Fixing obviously wrong data

Posted: Wed 19 Jan 2011 9:41 pm
by daj
A new feature to help you recalculate all your records is being tested in the beta version at the moment.

However, for now, you could revert back to an older backup of Cumulus data -- it keeps that last ten copies of the data files so hopefully you have not restarted Cumulus more than 10 times since the error. Copy the contents of one of the folders (based on the date/time) under the 'backup' folder and copy the files back to the Cumulus folder (with Cumulus closed down). Restart Cumulus.

Alternatively, and for speed if you just need to fix an all time record, use notepad to read the alltime.ini file from a backup folder and use the data to Edit the Records within Cumulus (Edit menu)

Re: Fixing obviously wrong data

Posted: Wed 19 Jan 2011 10:02 pm
by steve
The only other thing to add is that changes to all-time records are tracked in alltimelog.txt, so you can also find previous records in there.

Re: Fixing obviously wrong data

Posted: Thu 20 Jan 2011 6:10 am
by permaculturegeek
Well because I'm only downloading data from the receiver to cumulus weekly (we won't be living at the weather station site for a year or two), the backup option is no use.

I've loaded a copy of the data as csv into a spreadsheet, which has made min() and max() functions easier after removing the errant data. Replacing the records with the previous records was easy (given there were only 800 entries), but of course it is also necessary to check if a record has been broken since 13th January (the date of the error).

A recalculate feature will be most welcome!

Richard