The B-TREE-P system assumes a data file does not contain items with null item identifiers. Null ids are a bad programming practice that inevitably leads to confusion. Although BTPINS and BTPDEL will correctly index items with null ids, BTPSEQ and some versions of BTPFIND use the convention of returning a null id when no key is found. Programs that use BTPSEQ and BTPFIND, such as browsers, are normally coded to treat null ids as error or "end of tree" conditions. If your data files have null ids, your browsers will probably become confused when they encounter those items, and prematurely stop operations such as scrolling or sequential file processing.
As a rule, you should always prevent items with null ids from being saved in your data files.