Tree Search Algorithm


Active Member

I want to write a UBE to print BOM. I understand that there is a base JDE UBE to do it but it has some limitations (pointer issues) when run on large volume of data.

Since I am new necessarily I want to code this in NER rather than C. Thus I miss the arrays and datastructures :) Well I can simulate the arrays through Table so no problem.

I simply can't just come up with a logic to search through the BOM given a Finshed Good as input. The BOM data is stored in table F3002 as one record per Parent and child link.

Basically what I am doing is trying to search a tree structure (BOM) and just can't come with a logic. In the school I have read about alogorithm for Tree search like Depth-First search (Searching up to the last level at one time) Breadth-First search (Searching all nodes at a level at a time) and I think I need to implement them.

Can any of you guys provide me some clue/hints/opinion/suggestion on what could be the logic to search for BOM (Tree).
Remember you just have Tables (Files) for your workspace (no other data structure like stack/arrays/link list)

Take it as challenge guys!!



Legendary Poster

what problems have you run into using the standard BOM UBEs?
How deep/wide were the Bills that caused the problems you ran into?
I'm very interested in knowing this as we have several customized BOM explosion UBEs based on the standard UBEs and BSFNs.

Back to your question I think that if sheer size/volume of data is forcing you to a custom solution you will find that attempting to use tables in place of stacks/arrays will be extremely slow.

My 2 cents,

Larry Jones
OneWorld XE, SP 15.1
HPUX 11, Oracle SE 8.1.6
Mfg, Distribution, Financials

P.S. Next time post items like this to the Devlopers Forum<P ID="edit"><FONT SIZE=-1>Edited by Larry_Jones on 10/9/01 09:11 AM.</FONT></P>