Build 960 install

[bnwrx]

Should we uninstall the month and year.ini files first?

[steve]

Delete them before you run the new build.

[Synewave]

Well done Steve and many thanks.

It seems to have fixed the problems for me at http://www.synewave.co.uk/gcsweather/thismonthyear.php

I'll now look at that dodgy rainfall record.

[nking]

Steve,

Please don't read this if you're feeling tired or if you've not got a glass of wine in your hand

Cumulus_2010-11-02_15-48-36.png

Cumulus_2010-11-02_15-50-01.png

Do we need to manually update the ini files or should it pick up the data automatically. From the first screen grab Cumulus clearly knows the data is there but from the second it hasn't been able to find the correct data e.g. highest temp. - sorry if I have missed some instruction.

Not relating to this build but I notice the snow data isn't showing from Jan this year. I last recall seeing it in Feb when I think it showed up in an xml file - I'm probably wrong. Anyway, it looks as though I have lost it since it is not in any data I can find. Out of interest where is it recorded?

I'm sure it wouldn't be easy but wondered why Cumulus couldn't use the function of calculating the averages and extremes period to propagate the new high and low data elements. The possibilities for historical information reporting seem so near......

Forgive my podering but I did warn you only to read this if you're feeling

[steve]

Do we need to manually update the ini files or should it pick up the data automatically. From the first screen grab Cumulus clearly knows the data is there but from the second it hasn't been able to find the correct data e.g. highest temp. - sorry if I have missed some instruction.

You presumably have the same corruption mentioned in two other threads. I'm beginning to get an idea of where to point the finger of blame for this

Not relating to this build but I notice the snow data isn't showing from Jan this year. I last recall seeing it in Feb when I think it showed up in an xml file - I'm probably wrong. Anyway, it looks as though I have lost it since it is not in any data I can find. Out of interest where is it recorded?

It's in the 'diary' - log.xml.

I'm sure it wouldn't be easy but wondered why Cumulus couldn't use the function of calculating the averages and extremes period to propagate the new high and low data elements. The possibilities for historical information reporting seem so near......

Sorry, I don't understand...

[nking]

I understand that my data may be corrupted and this is why the reporting for “This Year Records Form” is incorrect but I then can’t understand how Cumulus is able to correctly calculate the “Averages and Extremes” for the same period – I guess it’s not from the same source data file?

I’ll try and elaborate on what I meant by historical reporting, unfortunately my written communication skills are not my forte.

The “averages and extremes” reporting which is an on demand user function produces excellent historical data. I’m not sure the most is made of this facility/data. If Cumulus automatically ran this function each month end and kept the data records then these could then be fed into comparative reports and/or graphs i.e. Average Monthly Temperature, Monthly Rainfall, Average Monthly Wind Speed / Wind Run, Number of Rain Days and so on. The monthly data could also be grabbed for uploading onto the website……

But then if I had wings I could fly

[steve]

I understand that my data may be corrupted and this is why the reporting for “This Year Records Form” is incorrect but I then can’t understand how Cumulus is able to correctly calculate the “Averages and Extremes” for the same period – I guess it’s not from the same source data file?

Because the new 'highs and lows' contain more items than the 'averages and extremes'. And it's in determining those extra items where it hits the erroneous extra separators on the end of the first lines in dayfile.txt which cause it to give up in a mixture of confusion and disgust.

The “averages and extremes” reporting which is an on demand user function produces excellent historical data. I’m not sure the most is made of this facility/data. If Cumulus automatically ran this function each month end and kept the data records

I'm in two minds about this sort of thing. The data is there in the log files in CSV format, so anyone wanting this sort of analysis could import the data into a spreadsheet or database and do it for themselves. On the other hand, the whole point of Cumulus is to provide information, so shouldn't it be doing this itself? I do intend to add NOAA monthly reports at some point, which is a similar sort of idea.

[SUBYDAZZ]

Should we uninstall the month and year.ini files first?

Is this only if you've been having some problems, or should we all do it regardless for versions after 958 up to 961?

[Synewave]

I've deleted them on every build I installed today and Cumulus has always recreated the files.

[nking]

CCumulusdatadayfile.png

By referring to the diags I found an error with the dayfile where the first 37 records finished with “,,” upon removing these I deleted the ini for month and year and reran Cumulus and to a large extent this improved the results. However, there are still anomalies within “This Year Records Form” and the diags give no clue to the reason so I'm accepting defeat and I guess the best option will be to amend the respective ini files manually – is my assumption correct?

[steve]

CCumulusdatadayfile.png

By referring to the diags I found an error with the dayfile where the first 37 records finished with “,,”

Yes, that was the problem that others have reported. What have you been using to edit the file?

However, there are still anomalies within “This Year Records Form” and the diags give no clue to the reason so I'm accepting defeat and I guess the best option will be to amend the respective ini files manually – is my assumption correct?

Well, it depends on what the anomalies are. If it doesn't encounter a problem in the file, the results should be correct.

[nking]

What have you been using to edit the file?

I now use notepad++, but it was probably notepad before that.

Well, it depends on what the anomalies are. If it doesn't encounter a problem in the file, the results should be correct.

I have used the editor picture as all the data iws on one screen

All time records editor_2010-11-03_14-09-36.png

ThisYearRecordsForm_2010-11-03_14-18-28.png

ThisYearRecordsForm_2010-11-03_14-22-04.png

ThisYearRecordsForm_2010-11-03_14-24-03.png

[steve]

I now use notepad++, but it was probably notepad before that.

I can't see that notepad[++] would have put those extra separators on the end of the lines that some of you have. I was thinking perhaps Excel...

I have used the editor picture as all the data iws on one screen

My note in the original announcement was obviously not clear enough, because no-one seems to have understood what I was trying to say. The import code only looks at dayfile.txt and today.ini. The items you have highlighted are (nearly) all recent additions to dayfile.txt, so there's nothing there for the import code to find. I didn't write the extra code that would have been needed to go through the all-time records on the off chance that there was something there that could have been added to the new displays - sorry.

The off-by-one on the lowest maximum date: could you have a look in your log files and see on which date that actually occurred? The new code takes the date from dayfile.txt and uses that, so you would expect that one to be correct. The (existing) all time records code does the check at the day rollover, which should be just after midnight. It subtracts an hour and uses that date. So they ought to be the same - the timestamp in dayfile.txt is 'now' minus one day, which should come out the same, just after midnight.

[mcrossley]

I can't see that notepad[++] would have put those extra separators on the end of the lines that some of you have. I was thinking perhaps Excel...

I see that file has the extra fields on later rows, so if edited in Excel and saved as a CSV, Excel will add the extra field separators for the lines with the 'missing' fields.

[nking]

I can't see that notepad[++] would have put those extra separators on the end of the lines that some of you have. I was thinking perhaps Excel...

You're probably right! Whilst my memory of Jan 10 is poor I expect I tried to be clever editing the day file using Excel

I didn't write the extra code that would have been needed to go through the all-time records on the off chance that there was something there that could have been added to the new displays - sorry.

The off-by-one on the lowest maximum date: could you have a look in your log files and see on which date that actually occurred?

First off, I think you did/do a great job for us all in making our lives easier. My concern was that my data may be suspect and that's why it wasn't working for me (I thought it was only me that wasn't getting it to do everything ). Manually entering the data is really no hardship.
Re the lowest max. it was first recorded 7th Jan 23:57