Yeah, I'm pretty sure this isnt needed

nt_summary_1 <- tbi$day %>% left_join(tbi$hh) %>% select(survey_year, person_id, day_id, day_weight, starts_with("no_travel")) %>% select(-contains("other_specify"))

nt_summary_2 <- tbi$day %>% select(survey_year, person_id, day_id, day_weight, starts_with("no_travel")) %>% select(-contains("other_specify"))

identical(nt_summary_1, nt_summary_2)

I think you want this

tbi_data_upcoded_copy <- copy(tbi_data_upcoded)

Otherwise tbi_data_upcoded_copy will act like a pointer to tbi_data_upcoded, and not like it's own dataset. If you're in data.table that is.