Recommendation: Give all your QVW files unique names. Do not, for example, maintain both;
Sales\Dashboard.qvw
Finance\Dashboard.qvw
Instead do:
Sales\SalesDashboard.qvw
Finance\FinanceDashboard.qvw
If you want to stop here at TLDR (Too Long Didn’t Read) just take the recommendation. Use unique names. For the geek crowd, read on,
The reason for preferring globally unique names is metadata. QlikView provides a number of different tools and facilities for scanning QVWs that provide useful (metadata) information for managing your pool of QV documents. Among them are:
– The lineage elements in a QVD that tell you what QVW created this QVD.
– QvsAdminDataProvider.dll that exposes information on user created server objects and ownership.
-QvServerObjectConnector.exe that exposes information on the Dimensions and Measures used in those server objects.
– A scanner you may write yourself (or one of my tools) that reads information from QVD or QVW files.
Each of those providers may use a different full path representation for the QVW file. These differences result in a different “identity” for the QVW and create difficulties in coalescing these very useful bits of information together to the common QVW file.
The lineage elements record the path at reload time, which is probably the server path, which does not align with the mapped drive you use in your scanner app.
QvsAdminDataProvider uses Access Point paths, which are URL paths like:
Dist/Sales/Dashboard.qvw
QvServerObjectConnector uses physical paths to the .shared file which are like:
X:\qlikvewserver\DistributedDocuments\Sales\Dashboard.qvw
These different path points of view make it difficult to tie the information together. If we can assume that the qvw name “Dashboard.qvw” is an identifying key, there is much insight available by combining these various metadata sources,
Bottom line. Give your QVW files globally unique name to get the most leverage of the metadata.
-Rob
There are still seats available for the San Francisco Master Summit for QlikView. Join us to learn about Advanced Scripting with me, Visualization Techniques with Bill Lay, Data Modeling with Barry Harmsen and Advanced Set Analysis with Oleg Troyansky. I’ll also be showing off some cool analysis of metadata such as what Dimensions and Expressions your users are using in their Server Objects.
