http://qs321.pair.com?node_id=438477

punkish has asked for the wisdom of the Perl Monks concerning the following question:

order monks,

I have, what sounds like, a common problem. I am creating a survey. A table holds the questions.

QUESTIONS ----------------- QUESTION_ID (pk) QUESTION_NUM (int) QUESTION (varchar) SHOW_THE_QUESTION (bit)

I have a 100 odd questions in already. Everything works. A colleague comes and adds 3 more questions, "Yo punk, please add this question after the current q 10, that question after the current q 59, etc." You get the picture. Now I have to go and re-order the darn QUESTION_NUM field. What a pain in the you-know-what.

So, I decided to implement the old FORTRAN, BASIC kinda system... provide a way to insert something in the stack so that the stack rearranges itself... allow the QUESTION_NUM field to accept floating points. So, I can add a new question, and enter 10.5 for its QUESTION_NUM, sort on the said field, and bingo.

Of course, then in the program I have to do some post-processing. The users won't understand what the heck q 10.5, or 10.725 (a result of many a new inserts) is. So, I have to renumber them after retrieving them from the db just for display purpose.

My problem doesn't end, however... some questions refer to other questions, "If you answered 'Yes' in q 20..." Of course, q 20 on the view side might QUESTION_NUM 19.25 in the table.

I am sure there are other pitfalls. In fact, just by checking off the SHOW_THE_QUESTION field I can set off the same problem.

Since this seems to be a fairly common problem, what are the creative ways that monks have solved or would advise solving this?

--

when small people start casting long shadows, it is time to go to bed