This is a feature of Publisher as the distribution task will be making the QVW available to your user group, and in most cases you wouldn't really want them looking at your script.
I cannot check for sure, but you might be able to just perform a reload task?, and get the publisher server to save the QVW into another folder with a new name instead of performing a 'distribution' on it.
Of cause this wouldn't allow you to use the loop and reduce features of publisher, but may get to the result your after.
I am interested in confirming the premise the script is always removed from distribution copies (copies as the end product of a reload task).
We have tasks set up in which the Source folder is also mounted published folder.
When the task reloads, the .QVW log shows a valid reload, but the script is still present, and is the only known copy of the .QVW.
When we adjust the task to direct the distribution copy to a location not re-pointing to itself, a second distribution copy is created and the script is removed.
When we explicity re-point the distribution copy back to itself, we get the warning message, and the reload removes the script.
So it is in cases where the 'Distribute:Distribute to Folder' path is left blank, that the task is performing a Reload, skipping the creation of a distribtion copy. Then the reloaded source copy is all that remains in the AccessPoint. (because the Source Copy folder has been mounted to Access Point)
Is this possible?
Yes, but about every other setting belonging to the distriubtion will not be made. If you don't distribute, you will not get access permissions set to anything else than what the file already had, you will not be able to grant creation of server objects to some users, you will not be able to set client preferences,... In short you will not set any meta data that controls properties and access rights of your published document.
Making Source and User folders point to the same location is strongly discouraged. But disabling distribution simply switches off various nice automation techniques in Publisher.
For some analogies, see here: Distribute to Folder vs Distribute to Server
This is something we're seeing also, as new users of Qlikview. We actually lost a couple of pretty complex scripts before we figured out what was happening.
Problem is, we cannot reload the data (or load new data) onto a QVW that's been put in the access point. This should be an automated function. I'm looking into Publisher now for answers.
I am using Publisher successfully without losing scripts, in a nutshell here is what I do:
- Develop QVW in Dev folder[s]
- Copy it to the Apps.Source folder, which is defined in the QMS as a Source Folder
- Publisher Task reloads from Apps.Source and distributes into the Apps.User folder which is defined as a Mounted Folder
So Dev & Apps.Source have script.
And Apps.User does not, which is what we want.
Would that work for you ?
If not then which bit trips you up ?
Best Regards, Bill
No, we definitely have the source and user folders set to different destinations. It's not overwriting the original folder, although we did have some admins who were overwriting source files with 'updated' versions (front end changes) based on QVWs from the user folder, thus no script. I've taken care of that.
My problem is that one of our distributed documents is losing data and returning empty charts and other errors from the Access point. I'm going to post that as a separate question.
If you set the source and distribution folder set to be the same destination will you get this behavior as the distributed copy (that never has a script) of the document will overwrite the source.
Starting from v11.20 SR4 you will get a warning if you set up a task where the source and distribution folder are the same to prevent tasks to run and overwrite the source document.