Most of the work for locating images by the ids used by the butler to locate the files has been done in DM-7494. KT has provided the following formulas for breaking a single number into the individual components.

field = ccdExposureId % 10000
camcol = (ccdExposureId//10000)%10
filterName = "ugriz"[(ccdExposureId//100000)%10]
run = ccdExposureId//1000000
patchY = coaddExposureId%(2**13)
patchX = (coaddExposureId//(2*13))%(2*13)
tract = coaddExposureId//(2**26)
patch = "%d,%d" % (patchX, patchY)

I gave incorrect formulae for coadds since we need to extract the filterName from our per-filter coadds as well. The code should be:

filterName = "ugriz"[coaddExposureId%8]

patchY = (coaddExposureId//8)%(2**13)

patchX = (coaddExposureId//(2**16))%(2**13)

tract = coaddExposureId//(2**29)

patch = "%d,%d" % (patchX, patchY)

Note that the code for coadds and that for calexps is different, and each can only be applied to its particular type of id.

Per John Gates, the issue may have been fully resolved already, just verify.

The work has been done in DM-7494 by jgates108 (GH tag: dax-F16-2016-11.pre2).