information flow in EA

[alexander komakula]

hi,
am trying to create a infomation flow connector and add a signal and realize it using automation..is it possible

[fwoolz]

IIRC, EA uses some database hackage to tie an information item to an information flow... I'll have to see if I can retrieve my recollections on this from the dusty vault in my head...

[fwoolz]

OK, here's what I found:

1) The Information Item is linked to its associated Information Flow connector by an entry in table t_xref in the EA database, as below:

Fields:
XrefID, Name, Type, Visibility, Namespace, Requirement, Constraint, Behavior, Partition, Description, Client, Link

Values:
{755E6BE2-3935-4251-8EFE-3FC343EC5327}, MOFProps, connector property, Public, , , , conveyed, 0, {153649D7-4576-43e3-AFC9-164C8E54F20F}, {35FF94D1-A91D-46ff-99E7-8DA48A27150E}, <none>      

The fields 'Description' and 'Client' are the ea_guids for the information item and the information flow, respectively.

2) If another connector (say, an association) realizes an information flow, the connector and the information flow are linked by an entry in table t_xref in the EA database, as below:

Fields:
XrefID, Name, Type, Visibility, Namespace, Requirement, Constraint, Behavior, Partition, Description, Client, Link

Values:
{8CB6EFC6-E0E0-4614-92E2-1895BAF4C3C5}, MOFProps, connector property, Public, , , , abstraction, 0, {35FF94D1-A91D-46ff-99E7-8DA48A27150E}, {DA2F4C80-6125-453b-AAF3-51B81B9DDD82}, <none>

The fields 'Description' and 'Client' are the ea_guids for the information flow and the association, respectively.

3) If you want to do this programmatically, it appears that you would need to add the necessary rows to table t_xref yourself, either by using the undocumented AI method Repository.Execute(<SQL statement to add a row>) (see http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1284736693/1#1 for more info) or by accessing the database directly using the database driver of your choice.

Word to Sparx - this outta be addressed in the API.

[alexander komakula]

Thanks for the information.but instead of directly updating table are there any api's for doing so

[Geert Bellekens]

Alexander,

If you read Freds reply carefully you will notice that the link to the information items is not exposed by the API, hence the complex database solution offered.

Geert

[Kris Nunes]

This is a super cool feature to have access to via the API.

I see it being used extensively in Data architecture when we need to map processes, data and systems (DFDs, CRUD,,,,)

Will this be possible in the upcoming releases of Sparx? If someone knows someone in Sparx, they need to point this missing API.

regards,
Kris NUNES

[RoyC]

Kris (and anyone else interested)

Please submit this as a Feature Request (or, if you prefer, a bug report) directly to Sparx Support using the links in the Support column at the foot of this page.

Thanks

[Guillaume]

Until information flows are supported via EA API, can't we use the SQLQuery method on the Repository?

[Dave_Bullet]

t_xref is the "black hole" table in EA.  If you can't find what you are looking for in your regular EA base tables, it's probably living in t_xref.  It allows Sparx to add features without worrying about having people apply DB patches to upgrade, support multiple different EA versions & backward compatibility.

A feature request would make sense to abstract the complexity of this table and also insulate future DB level changes by Sparx.

[Colin Richardson]

The PDATA columns also hide a wealth of information...

Best bet for this sort of stuff is turn on SQLProfiler and do it in EA and check your trace.