hi,everyone
i am new on bo field,but i meet a great doubt,i don’t split the chasm concept.
1) B(n:1)-A-(1:n)C
i think it’s real chasm,the invaild fault need two contents to slove,but i meet thus a strange problem,if the first "1"in table a(b.id=a.id) is different with the second "1"in table a(a.name=c.name),how to slove it?
2) B(1:n)-A-(n:1)C
i think it’s not real chasm.here,even if i make two contents,when i choise the most simple sample to examine(table a and table b),the result is also wrong,i don’t know why?
i very appreciate your help.
I don’t understand you problem completely.
The cardinalities you have changed in scenario 1) and 2) have no influence.
The only thing that counts is, how your contexts are set.
Carinalities are only used as info and for detecting contexts automaticaly. Cardinalties don’t change the contexts automaticaly…
The questions are:
What is(are) the ‘fact-table(s)’ and ‘dimension-table(s)’? Scenario 1) is the 2 context with 2 fact-tables (B and C), scenario 2) is the 1 context with only 1 fact-table (A). You have to choose what you want…
Are the joins between the table complete: all relevant columns joined?
hello,JdVries
thank you for your reply.i don’t explain clearly because i view others univers.1) and 2) are my created sample.
in my problem,A is Fact-table,B and C are Dimetion-table.under this condition,i want to know which is chasm model(1) or 2)).
in other word,i don’t search the detail of chasm trap from all post,but i only know what different status is belong to chasm?like the page"Fan Trap Question"
thanks
You won’t have a chasm trap with a single fact table, as the fact is the “many” rather than the “one” and it is in the middle of the joins, as in:
D1 --< F1 >-- D2
There is no chasm there. If you had more than one fact table, and you share dimension tables:
F2 >-- D1 --< F2
Now you have a chasm, as you cannot combine rows from two fact tables simply based on their relationship to the dimension table. You must have separate queries for this to work.
hi,Dave Rathbun/JdVries,
thank you for your helpful information.ok,now i can continue my question.for your say:“You won’t have a chasm trap with a single fact table, as the fact is the “many” rather than the “one” and it is in the middle of the joins, as in: D1 --< F1 >-- D2 There is no chasm there.”
follow my links:
D1 --< F1 (D1.department_id=F1.departmentid(+));
F1 >-- D2 (F1.code(+)=D2.toolcode);
When i use the special function"Detect contents",the system prompt me add the content among of three table(D1,F1,D2).if i need agree but it’s not chasm?
after agreement,i can’t use it(D1.department_id,D2.toolcode,F1.measure) in Bo at the same time,the prompt is “Exception: DBD, ORA-01417:
State: N/A”,so i can only chooice (D1.department_id and F1.measure) or (D2.toolcode and F1.measure),i don’t know?
thanks
vince
First, are the (+) symbols for outer joins? Don’t use 'em. In fact, that’s the reason for your Oracle exception, you have outer joins that contradict each other.
(BOB member since 2006-02-06)
(BOB member since 2002-06-06)