Hi there,
Multiple transaction tables, that don't have a 1:1 or at least 1:M relation, defined by some sort of a unique key, should NOT be joined, or else they will generate a lot of duplicate values - because Qlik will generate a row for each combination of matching keys from both tables.
To give you an example - if the only matching field is Country, and you have 100 transactions in Table 1 and 50 transactions in Table 2, the resulting table after the Join will contain 100 x 50 = 5,000 rows. This is certainly not what you want to see.
Typically, multiple transaction tables that have some common dimensions, are either concatenated into a single Concatenated Fact, or linked together with a Link Table. These are two most commonly used models in Qlik Data Modeling. You can look up blog articles or books (including mine) that explain these methodologies.
Allow me to invite you to the upcoming Masters Summit for Qlik event in Vienna, Austria on September 30th. We will be teaching many advanced development techniques, and this will be one of the topics - advanced data modeling, and how to deal with multiple transaction tables.
Cheers,
