Hello all,

When I submit a set of regions using Jigsaw the names of the output files change depending on which regions I have active. for example, if I have six defined regions, and turn the first five off, then the last region, number six will save its files named as region 0, so “XXXX_region_0_.0001”. I would expect that output filename would be locked to the region number.

Best,

J.

Thanks for looking Bobo.

J.

I took a second look…

This is tricky, because the Index is incremented not only for the Regions you define, but also for the Sub-Regions of the Regions.
So for example if Region 0 has sub-divisions set to 4 and 3, you will get Region Index 0 to 11 for that Region alone, and Region 1 will have index 12… So if a change to the sub-divisions of a single region would shift all following regions’ indices around, disabling a Region and letting it change the index isn’t that much worse.

The Index is currently used in both the JOB property naming (OutputFilename%Tile%=…) and in the actual filename.
We could modify the filename to include the subdivisions too, so it would be XXXX_region_0_1x1_.0001, so for the above example, we would get file names like

XXXX_region_0_1x1_.0001 XXXX_region_0_2x1_.0001 XXXX_region_0_3x1_.0001 XXXX_region_0_4x1_.0001 XXXX_region_0_1x2_.0001 XXXX_region_0_2x2_.0001 XXXX_region_0_3x2_.0001 ... XXXX_region_0_4x3_.0001
to ensure the region index stays even if the sub-divisions change.
We could go even further by adding the max. sub-divisions to the name, e.g.

XXXX_region_0_1x1of4x3_.0001 so if you change the Subdivs to 5x2, the filename would be uniquely XXXX_region_0_1x1of5x2_.0001 and won’t collide with a previous output using the 4x3 subdivisions…

Thoughts? Opinions?
The change to the code would be significant (the region is being defined for every output type in dozens of places in the code), so I don’t want to go and make changes that wouldn’t be welcome.

Also, we could make the region naming optional, allowing for the current behavior, and for both new behaviors outlined above to make sure we don’t break existing pipelines and allow for further naming conventions. Maybe even add $tokens to define the region naming, e.g. “region$index_” , “jigsaw$regionID_”, “reg$regionID_$tilesXx$tilesY_”, “myregion$regionID_$tilesXx$tilesYof$maxTilesXx$maxTilesY_” and so on…

But that would become a whole new project…

Hi Bobo,

Thanks for the additional detail. It isn’t a stop work problem for me, and as I am the only person who noticed I doubt it should really have that much priority.

Best,

Josh.