semi-structured data

Semi-structured data is data that has not been organized into a specialized repository, such as a database, but that nevertheless has associated information, such as metadata, that makes it more amenable to processing than raw data.

The difference between structured data, unstructured data and semi-structured data:
Unstructured data has not been organized into a format that makes it easier to access and process. In reality, very little data is completely unstructured. Even things that are often considered unstructured data, such as documents and images, are structured to some extent. Structured data is basically the opposite of unstructured: It has been reformatted and its elements organized into a data structure so that elements can be addressed, organized and accessed in various combinations to make better use of the information. Semi-structured data lies somewhere between the two. It is not organized in a complex manner that makes sophisticated access and analysis possible; however, it may have information associated with it, such as metadata tagging, that allows elements contained to be addressed.

Here's an example: A Word document is generally considered to be unstructured data. However, you can add metadata tags in the form of keywords and other metadata that represent the document content and make it easier for that document to be found when people search for those terms -- the data is now semi-structured. Nevertheless, the document still lacks the complex organization of the database, so falls short of being fully structured data.

In reality, there is considerable overlap between the boundaries of the three categories, which are sometimes described collectively as the data continuum.