The “data dictionary” in a database tool (eg. Oracle or Informatica) is simply a repository of metadata.

Metadata is however data about the data.  The following are examples of metadata:-

• The structure of each of the Dimension and Fact tables, including the Primary Key, Natural (or business) keys and any hierarchies implicit in the data.
• For each Fact or Dimension table the record of the number of rows loaded, when they were last loaded, and indications of the data quality
• Some systems capture data at different time points from several systems, and use metadata to “release” a consistent set of data.  Equally the metadata is available for query, so users may see for example that the SALES records from Northern Division are correct to last night, but Southern Division are loaded to last week.  This means the data is correct within Division, but cannot be calculated correctly to date.  Ie.  Total sales are only valid up to “last week”.
• Every Dimension table may hold “groups” of columns.  For example a SALES_REGION_ID, SALES_REGION_NAME, REGION_LONG_NAME, REGION_POPULATION.  Some systems hold metadata about the relationships between these, so they are always held consistently (eg. Region 24 cannot have the name for Region 25).

 
In short the data in the data dictionary is an example of Metadata.  However, most people see “metadata” as additional information over and above the basic structure – information which adds value to the data in the warehouse.