...
Run Analysis on Trans List with Identifiers. You will get the same result as from the previous run. The flag Allow Partial Set Processing does not change the results of a model. It is useful is just such cases as this, when you use a Calculate by Set to generate a unique identifier for every record in a data set, or when you want to use the cache() function (see the next exercise) to calculate a set of running totals – and some of your candidate sets may be large.
In this exercise, you are not using a grouping at all on the input pipe, and so you will use a single candidate set to handle all records in the input data set. Setting this flag means that CenterView PhixFlow will not attempt to load the full input data set before starting to process it.
This flag is not suitable when you wish to calculate group totals – sum, max, min, etc. – for each key value, since you may not have all records available when calculating these group totals.
...
- Add an attribute to Bill Running totals with Name BillsRunningTotal
- Use the cache() function to calculate a running total for the amount billed so far on each account. Remember to use, as a key in the cache function, a different key from that used in the expression to calculate the number of bills so far (if you use the same key for both, the values will clash). You can use the ACCOUNT_NUM field again, but this time concatenated with a fixed plain text string, e.g. _out.ACCOUNT_NUM+"-RunTotal"
...