process configuration xml ID value restrictions, impacts on Insight ETL, etc. restrictions?
As we start to get further into our deployment lifecycle, we have multiple projects sharing one field in a workitem,
an enum.
the VALUES are all different, but the IDs (as generated) are all the same
com.ibm.team.enum.i2, i4, i6, i8, etc..
thus we think we will/can have collisions of these attributes when reporting across multiple projects in the data arehouse. (defects by feature by product)..
so, we want to make the IDs specific to each project, and considered using the displayed NAME for the tail of the ID.
but of course, nothing is simple, many of the names contain special characters,
/ & : + - % ... spaces..
the first 3 look to be restricted for sure. is there a documented set of allowed characters in the process config xml for element IDs? anyone tested the exotic when the data gets ETL'd by RTC jobs or Insight jobs?
thanks
an enum.
the VALUES are all different, but the IDs (as generated) are all the same
com.ibm.team.enum.i2, i4, i6, i8, etc..
thus we think we will/can have collisions of these attributes when reporting across multiple projects in the data arehouse. (defects by feature by product)..
so, we want to make the IDs specific to each project, and considered using the displayed NAME for the tail of the ID.
but of course, nothing is simple, many of the names contain special characters,
/ & : + - % ... spaces..
the first 3 look to be restricted for sure. is there a documented set of allowed characters in the process config xml for element IDs? anyone tested the exotic when the data gets ETL'd by RTC jobs or Insight jobs?
thanks
Comments
Kevin Ramer
Mar 18 '13, 4:28 p.m.This is an aside regarding the ID in enumeration types:
There are several gotchas with respect to altering the value of id on existing enumerations. You just can't alter the id w/o taking precautions.
To change use of any particular id, mark the one going away with some eye-catching string in the name attribute, create a new entry with desired id and name. Reassign all work items whose attribute has the "old" value. Delete the old value from the process configuration. Lather, rinse, repeat.
However, you could have a simplified scheme for the creation of the id values. For example paX.enumeration.literal.l1 Were X would be maybe numeric if projects few and would use 1 for the first presented (web ui) in the projects list.
The ID in the database table I think is 250 chars.
sam detweiler
Mar 18 '13, 4:56 p.m.yes, they are a lot of 'fun'. we will also be using the 4.x process sharing fragments for this too. I hate the idea of anonymous IDs.. you can't tell what you are looking at. without some special decoder ring at least. !