No result in Results tab

[kogexo]

Hello all,

Thank you very much for this bright and clean new version which is very promising.
I have some trouble using the Results tab. My runs seem to have worked but I can't find a way to visualize the results. There is no scenario displayed and when I try to Refresh, it says "0 scenario files will be read" and then (obviously) "Selected files do not contain any Tag(s) which can be read with veda". Do you have any idea of what I did wrong? Or what did I forget to select?

I guess you can still use VEDA BE to read VD files (but the cases appear with the date - SCENARIONAME_1709 - which is quite annoying for now).
It appears that I have weird results in my model from the VEDA runs. Might be that I didn't do all "different things" well. My model is quite small but I don't think I understand all changes required to use the model in new VEDA (for example, I don't know how to do a MIG table).

Thank you in advance for your help, and thank you for this nice development,
Matthieu

[AKanudia]

Glad that you like Veda2.0.

Set definitions are expected in the model folder now, in Excel format. The message "0 scenario files will be read" simply means that this file was not found. You can use the utility "Veda_SnT to Excel migration.xlsm", which comes with the installation (in Veda folder), to migrate your sets and table definitions from the old Veda_SnT.MDB.

You must have a UPD table in SysSettings that declares default interpolation options. Just rename it to ~TFM_MIG and change Year col to Year2.

Previous versions required the use of "Refresh" in Results, to get new scenarios. But the current version will recognize additional runs and give you the option to include them.

[kogexo]

Thank you Amit, I will test all that!

In the mean time, I have another problem, it seems like my BY processes are not rightly read. All regions have for example the same Stock than first region read (that is normally the main parameter changing).
Do you have any idea what could be the reason this time?
It seems weird to me that there might be an error in VT files as first region appears to be well read and other files are identical except for some values that depend on the title of the file. Regions are also well defined according to VEDA2 browser. I've tried to check Demo models with two regions and nothing seem different to me. To conclude, I'm quite lost

Thank you again!

[AKanudia]

Please post two screen shots:
1. the Excel table
2. view of the output in Browse

[kogexo]

Here they are

[Antti-L]

I assume the two screenshots showing process characterizations are for two regions?  If so, you could try specifying the region in the ~FI_T header:  ~FI_T: BDX01  for region BDX01, ~FI_T: BDX02 for region BDX02, etc.  I have used this method, and it has worked well.

[kogexo]

It doesn't appear to have any effect but thanks for the tip, it's way clearer now. Can I use a formula in the tag's definition? I still have the same PRC_RESID in all regions :S
I don't understand how it is possible When I double click from the Browse view on a value, it opens the right Excel sheet and shows the right value...which is different from what VEDA2 saved. Any other idea? :S

EDIT
Oh, I tried to put some values directly without using formula depending on the sheet name and it appears to have been rightly read by VEDA. Is it possible that something changed in VEDA2 that makes formula depending on filename not working anymore? Or do you have any idea of the reason that causes this problem?
I might be the only one to use such approach to fill my VT files since I have quite a lot of similar regions.

[AKanudia]

great, i think we are converging. Is calculation set to automatic?

I have never used sheetname based formulae, but never had any trouble with any other kind, including ones based on file name.

[Antti-L]

@Amit: I am using filename-based calculations regularly in BY-templates.   I hope that should still work under VEDA2.0?

[AKanudia]

It should

[kogexo]

great, i think we are converging. Is calculation set to automatic?

Do you mean the Excel option in File, Options, Formula, Calculation Options? It was already set to Automatic unfortunately.

Btw, my Excel is in French, I hope it doesn't change anything.

[AKanudia]

When you close the file, are the formulas showing what you need, or do you expect them to get refreshed while Veda is reading them?

Old Veda used to refresh files while reading, and Veda2.0 reads the file as it was saved last. We can find ways to refresh though, if deemed important.

Of course, files *are* refreshed in cases when Veda writes into them, like FILL tables and parametric scenarios.

[kogexo]

Before giving anymore insight, I'll just explain how I proceed:

• I modifiy BDX01 file
  


• I close it
  


• I use a batch script to copy it to BDX02, BDX03...until BDX19
  

I thought that it would be enough as each time I tried to open one of BDX02-BDX19 file, the right values were displayed. I just tried to read BDX05 (for example) excel file with another system (pandas lib from python), it appeared indeed that the value from BDX01 was stored and not BDX04 value. So you are right Amit, the refresh has not been done (obviously as Excel didn't open the file and VEDA2 doesn't do any refresh anymore). So maybe a refresh is necessary (but could it be done from batch or from my side automatically - without opening every single file obviously).

Btw, just by curiosity, I opened a file from my "old" model from VEDA before and after syncing it and the same value appears. So that means VEDA 1.0 still refresh the formulas but doesn't save them.

EDIT
I just saw your last EDIT, can we implement an empty FILL tab in VT files? I believe no but I can be wrong. That could fix my issue

[AKanudia]

great, so the mystery has been solved. The best option I see is for you to open and save each file, as an Excel object, after it is copied. However, I did not find any easy way to do that via Python, the last time I tried.

Veda2.0 opens files as Excel objects to search for tags and establish the range to be read for each tag, and closes them without saving. Then they are read as XML via Exceldatareader in a second pass. This approach has resulted in a major performance improvement and it is far more robust. VT files don't support FILL tables.

I would really not like to refresh files each time they are read, as they will all have a new timestamp, and that would be very inconvenient if you are using any versioning system on your model folders.

Old Veda makes temp copies of files before reading them. Each file is read in via an Excel object, which makes Excel practically unusable during the entire Sync process.

By the way, generally I have found it better to include multiple regions in single workbooks (super regions) when there are so many of them. See how the JRC_EU_TIMES is set up: https://github.com/KanORS-E4SMA/EU_TIMES_Veda2.0.git

[Antti-L]

Thanks Amit for the additional info.

In my case, my BY-master files DO NOT need refreshing, as my files are saved with their target name under Excel (and so are getting calculated accordingly). And I would also not want any unnecessary additional refreshing to done by VEDA. So, I am happy as it is now.