# Dimension

Statistical concept used in combination with other statistical concepts to identify a statistical series or individual observations.

A Dimension is a Component of an SDMX Data Structure Definition (DSD).

A Data Structure Definition comprises of 1-n Dimensions, each Dimension defines what its valid Representation is, and in the dataset a value must be reported against each Dimension to form the Unique Key.

__Example Dataset__

FREQ | REF_AREA | INDICATOR | TIME_PERIOD | OBS_VALUE |
---|---|---|---|---|

A | UK | H1B2 | 2002 | 12.3 |

A | FR | H1B2 | 2002 | 12.4 |

A | DE | H1B2 | 2002 | 12.5 |

In the above example, the Dimensions are FREQ, REF_AREA, INDICATOR, and TIME_PERIOD (which acts a special dimension to describe Time). Each row has a unique key based on the combination of these Dimension values. SDMX uses the term Series Key which is the combined values of the Dimensions excluding time (A.UK.H1B2) and Observation Key which is the Series Key including Time (A.UK.H1B2.2002). The Series Key is used to uniquely identify a single series in a time series dataset, which may comprise of multiple rows which iterate over time. The observation key will always uniquely identify a single observation (row) in a dataset. In a non-time series dataset the series key and observation key will be the same.

Dimensions differ from Dataset Attributes in that Dimensions are all required to uniquely identify a series. Attributes are not included in the unique identification of data, and can even be marked as optional.