Recent content by Golan

Just a thought: I have worked with Perl for ETL, but I considered it a wrong technical choice. When the data are big, Perl does not meet the performance of a binary program. Ok, you may say that the IO stuff is nearly as good in performance as in C, for instance, but when you speak of millions...

nlim, It worked: for a specific month I now know the customers that didn’t buy anything. For R1 I set “Metric Join Type” to “Outer” and for “Attribute Join Type” I used “Preserve common elements of lookup and final pass result table”. What’s next? Thank you!

I think the logic is ok, but when I use SUM instead of COUNT (that is, when I check the balance of the ‘account’ of each customer instead of counting the number of records for the sales fact for each customer), I can’t get the filter “Is null” or “Exactly 0” to work. It returns no data at all...

I keep hitting the same problem many times. I can’t find non buying customers unless I put a month attribute in columns, so that, together with the outer join setting of the count order metric, it show nulls (or zeroes if configured to) where the customer had no order in a specific month. But...

We have accomplished this by creating a lookup table where we have columns such as project’s name, timestamp of loading, and some attributes of our time hierarchy. Then we created an attribute called ‘reference date’ in each project, and a filter with an applysimple clause that selects the...