You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello, I'm currently trying to work with table instead of microdata with Tau-argus, I work for French NSI.
I have an issue about non-additivity, Tau-Argus says that the table is non-additive but in the non additivity report the Total is correct and at the end of the report it shows "Difference : 0". I did manual checkings and the total for value and frequency is correct.
I first thougth about decimals problems because there was an older issue about this, but with integers the problem remains.
When I select "Compute incorrect totals" instead of "Do not allow non-additivity" it works as expected and the Total is correct.
When I select "Allow non-additivity", the Total is missing and I only get a "-" .
If I remove the Total from the table, I still get the additivity error.
I can work with "Compute incorrect totals" but if there's an additivity problem I need to check the Tau Argus output to be sure that the data are correct.
My second point is about the "write batch file" option, here is the batch from Tau-Argus
Tau-Argus can't use it and here is the batch I wrote to make it work
The safety rules are missing but they were specified.
Is the write batch option reliable ? I can work without it but it can be useful for me.
This is not an high priority issue, I just want to have more information about the way Tau Argus handles tables.
Do I need to remove Totals/subtotals from the table or is it better to keep them to be sure that the output by Tau Argus are equals to mine.
Sincerely
The text was updated successfully, but these errors were encountered:
Next time please only ask one question / report one bug per issue. That way it is easier for others to search issues and it is clearer to the developers what is fixed and what still needs to be fixed ;-)
Question 1: Could you provide a minimal example that reproduces this behaviour? Without any data it is impossible to look for the cause.
Question 2: The write batch option was intended to provide a start for a batch file. You usually need to check the batch file, and if needed adjust it. I'll have a closer look later on how this could be improved. At least the used safety rule (if any) should be added.
Question/remark 3: you should provide totals, as long as they are from an additive table ;-)
Tau-Argus
4.2.0b5
Windows
Hello, I'm currently trying to work with table instead of microdata with Tau-argus, I work for French NSI.
I have an issue about non-additivity, Tau-Argus says that the table is non-additive but in the non additivity report the Total is correct and at the end of the report it shows "Difference : 0". I did manual checkings and the total for value and frequency is correct.
I first thougth about decimals problems because there was an older issue about this, but with integers the problem remains.
When I select "Compute incorrect totals" instead of "Do not allow non-additivity" it works as expected and the Total is correct.
When I select "Allow non-additivity", the Total is missing and I only get a "-" .
If I remove the Total from the table, I still get the additivity error.
I can work with "Compute incorrect totals" but if there's an additivity problem I need to check the Tau Argus output to be sure that the data are correct.
My second point is about the "write batch file" option, here is the batch from Tau-Argus
Tau-Argus can't use it and here is the batch I wrote to make it work
The safety rules are missing but they were specified.
Is the write batch option reliable ? I can work without it but it can be useful for me.
This is not an high priority issue, I just want to have more information about the way Tau Argus handles tables.
Do I need to remove Totals/subtotals from the table or is it better to keep them to be sure that the output by Tau Argus are equals to mine.
Sincerely
The text was updated successfully, but these errors were encountered: