Equal to Neo-Assyrian zarāpu, but that is attested only in that dialect and in identical context and would thus appear to be an Aramaic loanword. Probably a by-form of srp, "to suck up." For the identical phonetic alternation in the same period compare zrpnt with cuneiform Sarpanit.
The equally rare zrp #1 may or may not be related.
Neither Lemaire's treatment in the ed.pr. or that of Andrew Gross, Continuity and Innovation...(Brill, 2008) pp. 65ff., satisfies.
The JBA example may be related or, as suggested by DJBA be connected rather with zrb.
Page refs. in other dictionaries:
DJBA: 422bAudo: 1:296