COADD_EXPTIME==0, FIBERSTATUS != 0, but ZWARN == 0

@deisenstein identified that some zcatalog-denali-cumulative.fits rows have COADD_EXPTIME==0, FIBERSTATUS != 0, but still get ZWARN == 0. The vast majority of COADD_EXPTIME == 0 targets are have ZWARN bit 9 "NODATA" set, but how do we end up with some FIBERSTATUS != 0 (bad fiber) still resulting in ZWARN == 0 despite COADD_EXPTIME==0?
```
import numpy as np
from astropy.table import Table
zcat = Table.read('zcatalog-denali-cumulative.fits')
mystery = (zcat['COADD_EXPTIME'] == 0) & (zcat['ZWARN'] == 0)
print(np.sum(mystery)) # 1331 strange targets
print(len(set(zcat['FIBER'][mystery]))) # coming from 12 fibers
print(set(zcat['FIBERSTATUS'][mystery])) # all with FIBERSTATUS = 65536 = 2**16 = BADFIBER
```
ok, so they are bad fibers, but it seems like that shouldn't result in ZWARN=0.

Start with the zcatalog-denali-cumulative.fits data and reproduce the issue using the NumPy and table-reading snippet in the report. Trace the pipeline logic that produces ZWARN, COADD_EXPTIME, and FIBERSTATUS for the 1,331 matching rows, then verify that the cause is explained and the affected bad-fiber cases receive consistent warning behavior.

Written by the indexing model from the issue text.