Hangup when Branching to new scenario

[gbyrom]

The application is "hanging" when branching to a new scenario.

I think this might be to do with the fact that my current scenario has a new electorate with meshblocks assigned to it (previously, the new electorate was already in the database, but didn't have any meshblocks assigned). Now, in the new scenario, it has to allocate those meshblocks to the new electorate - but perhaps can't do that for some reason - possibly because it wasn't expecting the new electorate ID that they've been assigned to? SYmptom is the "Branching scenario" dialog box stays on-screen, and is "not responding". Previously this dialog box would go away quickly, and the branching process then ran after that.

[gbyrom]

OK, this may not be quite as much of a problem as I first thought. I left it runnig overnight and it DID eventually branch to a new scenario. No idea what time it completed!

[gbyrom]

Something strange has happened here. All of a sudden, branching to a new scenario takes no time at all. I haven't changed anything in the plugin, although I have edited the database to remove a whole load of old scenarios that were no longer needed, but I doubt that would've done it. In addition, it doesn't seem to switch to the new scenario any more. It did before! The new scenario is certainly created but the scenario_id in the Electorates table is still that of the scenario you're branching FROM.

[gbyrom]

OK more info here as I've done some testing:

• Branching to a new scenario from a scenario in which there's a new electorate, takes no time at all. THe database lines all seem to be correctly set up in the scenarios table and in the meshblock-electorates table. However, the electorates table still indicates the current scenario_id is the previous scenario (even though the name of the new scenario is displayed in the Redistricting Dock). This MIGHT be expected behaviour - I wasn't previously keeping an eye on the scenario_id in the electorates table; but I'm assuming it should change to the new scenario_id.
• Branching to a new scenario from the original (start) scenario, WITHOUT a new electorate in it, causes the application to hang up for about 20 minutes. Eventually it appears to have branched, however the same situation as above i.e. the scenario_id in the electorates table is not updated. The non-updating of the scenario_id in the electotrates table may not be an issue, but it would seem logical that it should be updated. All previous times we have branched to new scenarios, it has worked fine, but I never checked this ID so I don't know if it was updated or not. If this is expected behaviour then the only issue here is the difference in time between branching instantly (if there's a new electorate in the scenario) and taking about 20 minutes for the process to complete (if there isn't). That does seem a bit strange! It might not be an issue technically - but 20 minutes is too long to wait.

[nyalldawson]

Hmm - I can't reproduce this one. I would suggest making a copy of the database, and then running a "vacuum" operation over it. (This can be done by right-clicking it in the QGIS browser window). Let me know if that improves the branching speed...