Saving modified .xdi has no effect on server

[Paul Segal]

This is interesting. It even gives a fix!

http://www-01.ibm.com/support/docview.w ... RXT&mync=R

Regards

[jim wood]

Good to know Paul thanks,

Jim.

[Martin Ryan]

That is interesting.

I've had this problem several times and the root cause for me was conflicting aliases. Worst case scenario (which came up again yesterday) the only solution was to delete all the aliases, save the dimension down and reload the aliases.

Martin

I'm sure this issue will continue to come up, so a couple of search terms: dimension not saving, xdi file failing

[Olivier]

I've had this problem several times and the root cause for me was conflicting aliases.

It is also th case when a new element is identical to the alias of an existing element, when uploading the xdi everything seems to happen correctly.
There is a failure of dimension update in the background and the only trace would be on the server log file. ( Duplicate alias found).

I found it quite un convenient as there is no prompt warning you that the dimension has not been updated from the dimension upload front end in excel.
( You will only get the promt mentioning that the dim already exists and will be overwritten ) .

It seems that only checking that the latest Xdi upload is reflected in the dim can guarantee the update has happen effectively and if not going to the log will give an hint of the reason why.
( whether it is write access in xdi folder or conflicting aliases...)

[Martin Ryan]

It is also th case when a new element is identical to the alias of an existing element, when uploading the xdi everything seems to happen correctly.

Yes, I was a bit unclear. This is what I meant by "conflicting aliases" - creating an element that is already in the dimension as an alias.

I'm fairly sure this a regression bug. At least I don't remember having come across before version 9. I think it used to say that the element already exists, or that there was a conflict with an existing alias. But it's possible I'm making that memory up, it's been a while since I've worked with version 8.

[Olivier]

I think it used to say that the element already existing, or that there was a conflict with an existing alias

Martin,

I think I have a similar blur memory from the 8.xxx version...
We use to get something on the same line as the message that comes up if you forget to define a NLevel element which you have inserted in a consolidation...