SPARX RFC - Tagged Value support for CSV

[mrf]

There have been a number of requests both on the forums, and through the official support channels for the CSV import/export process to support tagged values.

Unfortunatley, representing this relational data in a two-dimensional format for a set of generic elements is not straight forward, and can be done in a variety of ways. As such we are requesting feedback from users as to what format they would expect tagged value data to be presented in.

This feedback will be evaluated through the design decision making process based on its merits. If you would like to participate your thoughts on this area, please send an email marked for my attention to [email protected].

[neil_albiston]

Has there been any progress on the import of tagged values via csv ?

[beginner]

I vote for error correction instead of any CSV "improvements".

b.

[Dave_Bullet]

I understand, since 2 different stereotypes could have different sets of tagged values (but some could be shared).
You could simply mandate that only "columns" in the CSV matching tagged values for a stereotype (nominated in the import process by a stereotype column selector) will be imported.  Using empty fields for invalid tagged values could be supported, yet mix and match in the one import.

For example, say you have 2 stereotypes - "application" and "server" (doesn't matter what the base UML classifier is).

Application has "IT support person" and "business function" as tagged values
Server has "IT support person" and "ip address" as tagged values

The import CSV file could look like this (1st row is column header):
element_name,stereotype,IT support person, business function, ip address
intranet,application,david teague,ERP,
w21223,server,john taylor,,192.0.0.1

The above is just a sparse matrix where common tagged values can be populated across stereotypes.

[Paolo F Cantoni]

I understand, since 2 different stereotypes could have different sets of tagged values (but some could be shared).
[size=18]...[/size]

You will need to also consider that the same stereotype may be in more than one MDG technology and have different sets of tagged values...

Paolo