The function dict_index_get_sys_col_pos() is performing unnecessarily complex computations, which only make some sense for DB_ROW_ID, which would exist either as the first field in the clustered index or as the last field in a secondary index (only when a DB_ROW_ID column is materialised). We should add accessor functions dict_index_t::db_trx_id() and dict_index_t::db_roll_ptr().

Also, there are some unnecessary calls to rec_get_offsets() when accessing the first field of a record, for accessing the columns DB_ROW_ID and FTS_DOC_ID.