On script tab 'FactSales - SCD Type2'
there is an Intervalmatch, which refers to field DateInvoiced:

// Temp
Temp_%KeyCustomer3:
Intervalmatch (DateInvoiced)
NoConcatenate Load
ValidFrom,
ValidTill
Resident Temp_%KeyCustomer2;

The resultset of this Intervalmatch contains 4 rows.

**

Searching backwards in the script the next occurence of
the field DateInvoiced is here :

Temp_%KeyCustomer1:
NoConcatenate Load
CustomerNumber as vCustomerNumber,
InvoiceNumber as vInvoiceNumber,
DateInvoiced
Resident FactSales_Source
Where RowId = $(i);

But: this only has a resultset of a single row !

**

Based on that this cannot be the source of the field
in used in the Intervalmatch.

To verify this I renamed the field:

Temp_%KeyCustomer1:
NoConcatenate Load
CustomerNumber as vCustomerNumber,
InvoiceNumber as vInvoiceNumber,
//DateInvoiced
DateInvoiced as DateInvoicedOM
Resident FactSales_Source
Where RowId = $(i);

Result does not change !

**

Searching further backwards in the script for
the field name leads to :

FactSales_Source:
LOAD
CustomerNumber,
InvoiceNumber,
DateInvoiced,
DateInvoiced as Invoiced.Date,
AmountInvoice
FROM $(vDataInput)\FactSales_Source.qvd (qvd);

This resultset contains 9 rows with 4 distinct DateInvoiced - values.

Question: why does Qlikview not refer to the next occurence of
the field backwards ?

I completely agree that you all have completed the challenge quite well with the date dimension table which reports well for at least for a month..The example helped me a lot to design my own dimension table so effectively and easily.safe cookware

Hey this a great information , i was looking for this information all over the net bust was not able to find , but now i can complete my work , thanks to you.vacation homes in orlando florida for rent

In dimension modelling, surrogate keys selection is very challenging tasks. But you nicely explained here how to select and simulate the application. free reverse cell phone lookup