Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
I have QlikView with Publisher installed on my server. I understand the SourceDocuments/ folder is needed for Publisher to set up tasks; and the Documents/ folder is for AccessPoint use.
When I do a document reload, I see a 12gb QVW under SourceDocuments and another 12gb under Documents folder. Why is this? How can I prevent this redundant storage? Without data, my QVW file is only 659kb..
It may seem wasteful for a simple Reload & Distribute, but I think there are good reasons for the design.
- There may be subsequent Reduce & Distribute task that expect data int the Source Doc,
- You may want to open the Source Doc for development, and would expect to find data without reloading.
Removing the SourceDoc data after Distribute would be an interesting product option, but it could create some problems. Loss of Always-one-Selected-Value and saved selections are two that come to mind.
If you want to store your SourceDocs without data, you could schedule a Reduce task that distributes to the SourceDocs folder. I think that may give you what you are thinking of.
-Rob
This is not a redundant step. This is the difference between task phase 1 = reload document and store in QVW, and phase 2 = distribute document to target (either a folder or an AP root folder or a server or whatever).
You cannot omit the source document (the published document hasn't got a load script anymore) or the published AP document (or you will not be able to open the document in the AP).
If disk space isn't scarce, leave the QVW documents as they are.
I have the server and publisher on the same machine. For distribution, it is set to Distribute to Folder and points to the /ProgramData/QlikTech/Documents folder. The QVW file is in the /ProgramData/QlikTech/SourceDocuments folder.
After a reload, the QVW file in SourceDocuments reports 12gb. Then it seems the file is copied by the distribution task to the Documents folder. This generates 2 - 12gb files, which I see as unnecessary. Why not have the "compiled" QVW file in Documents folder and the QVW without data in SourceDocuments folder? My original QVW without data is only 659kb, and this is put in SourceDocuments and the task is then run.
It may seem wasteful for a simple Reload & Distribute, but I think there are good reasons for the design.
- There may be subsequent Reduce & Distribute task that expect data int the Source Doc,
- You may want to open the Source Doc for development, and would expect to find data without reloading.
Removing the SourceDoc data after Distribute would be an interesting product option, but it could create some problems. Loss of Always-one-Selected-Value and saved selections are two that come to mind.
If you want to store your SourceDocs without data, you could schedule a Reduce task that distributes to the SourceDocs folder. I think that may give you what you are thinking of.
-Rob
In addition to what Rob suggests about distributing to the Source Documents folder, I think it would be useful to disable script removal in the distributed empty document in the Source documents folder (trick is documented in the community somewhere)
Or your Source Document will end up with a script consisting of a single comment. Which is probably not what you want.