dataset concept
You can define datasets and combine them to supersets. This mechanism lets you also define parallel and overlapping hierarchies of nested datasets.
Definition of a hierarchies is separated from the definition of dataset values for a concrete topic. So you can define a dataset grouping for regions, mapping continents, countries, states and departments and apply this grouping to define and combine sets of e.g. cities, person names or products.
We will apply the mechanism here for cities in geographical regions. You can find the example files in the distribution's directory demo/dataset/
A dataset is identified by a code. For a country, its ISO code is an appropriate choice, but you are free to define and choose what is useful for your application.
Assume you wanted to process some american countries: US (USA), CA (Canada), MX (Mexico), BR (Brazil), AR (Argentina)
You could group them geographically (North America vs. South America) or by language (Latin America vs. Anglo America). You could do both in parallel by defining area sets in a file area.set.properties :
| latin_america=MX,BR,AR anglo_america=US,CA north_america=US,CA,MX south_america=BR,AR america=north_america,south_america |
cities_US.csv:
| San Francisco Los Angeles New York |
cities_CA.csv:
| Ottawa Toronto |
cities_MX.csv
| Mexico Villahermosa |
cities_BR.csv
| Sao Pãolo Brasilia |
cities_AR.csv
| Buenos Aires Rosario |
You can now use this setup to generate city names for any of the specified regions. For north american cities you could specify
<echo message="north american cities:" /> |
and generate the output:
north american cities:
city[name=Mexico]
city[name=Los Angeles]
city[name=San Francisco]
city[name=New York]
city[name=Villahermosa]
city[name=Ottawa]
city[name=Toronto]
