In DataCleaner all sources of data are called 'datastores'. This concept covers both sources that are read/parsed locally and those that are 'connected to', eg. databases and applications. Some datastores can also be written to, for instance relational databases.
DataCleaner uses the Apache MetaModel framework for data access. From DataCleaner's perspective, Apache MetaModel provides a number of features:
A common way of interacting with different datastores.
A programmatic query syntax that abstracts away database-specific SQL dialects, and that is usable also for non-SQL oriented datastores (files etc.).
Out-of-the-box connectivity to a lot of sources, eg. CSV files, relational databases, Excel spreadsheets and a lot more.
A framework for modelling new sources using the same common model.
DataCleaners datastore model is also extensible in the way that you can yourself implement new datastores in order to hook up DataCleaner to legacy systems, application interfaces and more. For more information refer to the Developer resources chapter.