The role of plasmoid drifts towards the tokamak low field side and associated rocket motion of solid fragments in the material assimilation of shattered pellet injection (SPI) is investigated. In particular, a trace neon SPI scheme, where a small amount of neon is added to the pellet, envisaged to suppress plasmoid drifts via stronger radiation in the plasmoid, is examined combining experiments on JET and modelling with the non-linear MHD code JOREK. In the JET experiments, pure deuterium () pellet or pellet with a neon atomic mixture ratio ranging from to is shattered and injected into / H-mode plasmas. The experiments demonstrate evident rocket motion in SPI discharges, in contrast to trace neon SPI cases. However, this may be related to another key observation that the cooling time, defined as the duration between the SPI arrival and dip of plasma current before the current spike, drops drastically from about with SPI to about with neon SPI, which may not leave enough time for the rocket motion to occur. JOREK simulations show that plasmoid drifts play an important role in the SPI discharge and discover that the rocket motion in fact facilitates its core material assimilation since the fragments reached the high field side before moving outward. Simulations of trace neon SPI discharges demonstrate that the key limiting factor of their core material assimilation is their short cooling time rather than the level of plasmoid drifts in these discharges.