Which data model is better in this case?

Hi all,

there's a problem we've faced while modeling data for reporting and would like to ask for help.

There's Cube1 with characteristics: store, date, pos terminal and key figures Sales@Retail, Sales@UOM

And Cube2 with chars: store, date and key figure Customer Flow.

In reports we need to get the Customer Flow value for each line in Cube1, but it doesn't exist for each store.

So we see now two ways of solving this:

1st one is - creating infoset with cube 1 and cube 2 joining them with store and date. Then creating a multicube with Cube1 and our new infoset and basing the report on that multicube.

This way we'll get the customer flow for each store where there is one, and get no customer flow for each store where there's none loaded.

2nd way is - creating infoset with cube1 and cube2 using the left join by store and date. This is kinda simplier, but SAP doesn't recommend to use left joins, and it doesn't seem to work well enough.

So which way is better?

Thanks in advance,

Andrey