# y's target object should be large when retrival is "main" # and small when retrieval is "worker". tar_option_set(retrieval = "main") options(clustermq.scheduler = "multiprocess") pipeline <- pipeline_init( list( tar_target( x, data.frame(x = runif(5e6)), deployment = "main", format = "fst" ), tar_target(y, x, format = "fst") ) ) cmq <- clustermq_init(pipeline) debug(cmq$run_worker) cmq$run() # The target object should be several MB. # pryr::object_size(target) # Run in the debugger. # nolint tar_destroy() tar_option_set(retrieval = "worker") pipeline <- pipeline_init( list( tar_target( x, data.frame(x = runif(5e6)), deployment = "main", format = "fst" ), tar_target(y, x, format = "fst") ) ) cmq <- clustermq_init(pipeline) debug(cmq$run_worker) cmq$run() # The target object should be a lot smaller. # pryr::object_size(target) # Run in the debugger. # nolint tar_destroy()