May be someone here can help to untwist my brain. I could not find a solution up to now.
I have 2 tables and want to get a cross join. Sample:
Data Feed 1 returns:
SQUARE 08 1234
CALC 08 5678
Data Feed 2 returns:
07
08
09
I want to construct the following output within a Data Feed:
SQUARE 07 0
SQUARE 08 1234
SQAURE 09 0
CALC 07 0
CALC 08 5678
CALC 08 0
I didn't manage to achieve this with the Combine or Concatenate operators. Any idea?
Hi Donald,
yes, my issue is the same than yours. I have a time range (e.g month) and I have transactions. But there are month, where no transaction happens.
My issue goes even further. My time range may be year, month, day, hour, minute, second. I can even handle that, see below. I can also handle the days, depending on the month. But when I want to differenciate between transaction from application A and application B, I came to the point, where I think I need to do a cross join.
My solution for one dimension (e.g. month and transaction only): I have made a manual data source with one entry for each month (01 - 12) and transactions set to 0. Then I use the "Combine Data Feeds" operator and join real project data with the manual data feed. Afterwards I use the "Aggregation" operator to use the maximum values for the transactions only. This means the month - transaction pair, where the transactions are 0, will be removed.
I think an answer may be provided here: http://www.ariscommunity.com/users/mp3monkey/2010-08-27-processing-data-columns-instead-rows
"Please keep in mind that the number of data feed columns, their name and data type must be known at build time. They cannot change at run time."
Donald Dillon on
Georg,
I have a similar situation.
We have projects started by different groups. Each project has a start date (rounded to month)
Not all groups started a project every month
The results we get now
x-axis=month
y-axis=count of projects
the problem is that if a month had no projects the x-axis condenses and simply 'hides the gap'.
We want to display all months in the range and then of course the number of projects will be 0. I envision doing this somehow by extracting the min and max months in the records, generating the records to fill the gap and to give each month a record value of 0.
then I will join where the records with real values will be supplemented by the records with values of 0 to dynamically create a 'full' data set to be presented.
Do you see the similarities? I would be interested in any solution you come up with as it may be applicable for my scenario.
Thanks
Donald