Example: Grouping Records by Key
This example shows you how to group records with the help of the group-by function, and also illustrates how to aggregate data. This example is accompanied by a demo mapping available at the following path: <Documents>\Altova\MapForce2022\MapForceExamples\GroupTemperaturesByYear.mfd. This mapping reads data from an XML file that contains a log of monthly temperatures, as illustrated in the code listing below:
The business requirement of this mapping is two-fold:
1.Group temperatures of each year together.
2.Find out the minimum, maximum, and the average temperature of each year.
To achieve the first goal, the mapping calls the group-by function. To achieve the second goal, it calls the min, max, and avg aggregation functions. All of these functions are MapForce built-in functions, and you can add them to any mapping by dragging them from the Libraries window, see How to Add a Function to the Mapping.
The way MapForce executes a mapping (and the recommended approach to start reading one) is by looking at the topmost item of the target component. In this example, an YearlyStats item will be created for each group returned by the group-by function. The group-by function takes as first argument all data items from the source and groups them by whatever is connected to the key input. Since the requirement is to group temperatures by year, the year must be obtained first. To achieve this, the substring-before function extracts the year part from the month attribute of each data element. Namely, it takes as argument the value of month and returns the part before the first occurrence of substr. As illustrated above, in this example, substr is set to the dash character; therefore, if given the value "2006-01", the function will return "2006".
Finally, the values of MinimumTemp, MaximumTemp, and AverageTemp are obtained by connecting these items with the respective aggregate functions: min, max, and avg. All three functions take as input the sequence of temperatures read from the source component. These functions do not need a parent-context argument, because they already work in the context of each group. In other words, there is a parent connection—from data to YearlyStats— which provides the context for each aggregation function to work on.
To preview the mapping output, click the Output tab. Notice that the number of groups coincides with the number of years obtained by reading the source file, for example:
|Note:||For simplicity, the code listings above contain less data than the actual input and output used by the demo mapping.|