This way is a few hundred times faster (on large datasets)...NOTE:the details are missing but it still identifies the key holes in the sequence, except beiging and end which are obvious...?

SELECT...