Database structure

[mnmog]

Hello,

At my company we're using several instances of testLink, one for the verification team, one for the validation team and other for automatic tests. To obtain a better view of all the results and see understand globally the coverage of each product, we thought of creating our own database, similar to the testLinks's but without some unnecessary tables, where we would dump some data from the several testLinks and create views over it, filtrating only the wanted parameters.

However, since the ids from the several instances of testLink would conflict, we thought about changing the primary keys from ids to a combination of other parameters, depending on the data. After we saw in more detail the database (by the way, you should update the documentation and provide and updated DB schema), we realized that you have testLink data implemented as a tree.

I can understand this structure and it makes sense for most of the cases (I use a similar one in the automatic tests framework I developed) but I found two issues:
- I had several problems while importing a lot of XML data (test cases and suites) to testLink. I believe the amount of tables and data this structure requires may have been the cause of the problem.
- This makes our solution practically impossible, since everything is linked by the ids.

I also have a question: if for example one test case has the id 1, the steps have the ids 2,3,4,5 and the next test case has the id 8, if I had four or five more steps to the first one, what happens to the ids and all the connections of the subsequent ones?

I really would like some feedback on this subject. Thanks in advance.

[fman]

>>(by the way, you should update the documentation and provide and updated DB schema)
is not a good way to start asking for help, at least IMHO.
Hope that you will help with this task, because seems you have studied the schema => you have produced the documentation.
do you have this info to be shared with community ?

Is you do some search on our mantis site you are going to find => http://mantis.testlink.org/view.php?id=4004
On postgres we have added FK, this has to be of help on understanding table relationships.

For each node type detailed in table node_types, the ID is generated from a UNIQUE ID POOL, the ID FIELD on nodes_hierarchy table.
Then if you do this:
create test case => get INTERNAL DB ID = 10
create 5 test steps
create test case => you are going to get ID 16

[mnmog]

>>(by the way, you should update the documentation and provide and updated DB schema)
is not a good way to start asking for help, at least IMHO.

Sorry, I didn't mean to be rude or not polite.

Hope that you will help with this task, because seems you have studied the schema => you have produced the documentation.
do you have this info to be shared with community ?

Not yet, I'm still studying the schema and understanding how everything works.

Is you do some search on our mantis site you are going to find => http://mantis.testlink.org/view.php?id=4004
On postgres we have added FK, this has to be of help on understanding table relationships.

Yes, this is going to be really useful, thanks!

For each node type detailed in table node_types, the ID is generated from a UNIQUE ID POOL, the ID FIELD on nodes_hierarchy table.
Then if you do this:
create test case => get INTERNAL DB ID = 10
create 5 test steps
create test case => you are going to get ID 16

There's something I still don't understand. If, after you create the second test case with the ID 16, you add more steps to the first test case (which should be with ID 16, 17 and so on), what happens to the second test case ID and all the data that comes after it?

Thank you for the help.

[fman]

>> what happens to the second test case ID and all the data that comes after it?
I'm not able to understand your question.
Just know that ID are out of your control => you can not do any logic on it => as Mulder said ' Trust no one'