We have some strict incremental refresh requirements together with large data volumes. The current approach is to do a binary load of the "overnight" model and then append the day's updates.
My query centres around the QVW that we are reloading. This is approx 1.5 GB in size,and the opening and saving/closing of this data file can take up to 2-3 minutes. This together with the Incremental extracts means the incremental boundaries are being encroached.
However, given that we are doing a binary data load as the mandatory first step, it is pointless that this file starts out with any data and potentially if we didn;t have data in the starting file we could knock 1 minute off the opening time. Is it possible to open a QVW without data from the QMC, as if doing it from the Command Line. I know there could be DOS command workarounds here but I need the Logs to be visible in the QMC and Distribution Service Logs, which I don;t believe could be delivered by this solution.
I was hoping the QV Publisher Source/ User Documents would help. and thought that the Source Document would just hold the source code whislt the User Document contained the Data. This doesn;t apper to be the case though, and both documents hold data after reload. In an ideal world the Source document would contain no data, enabling rapid opening, it would distribute the saved data to User Documents, but the Source DOcument wouldn;t be saved again hopefully saving some time.
Has anyone encountered this before and can offer a suggestion?