My first Publication Agile-Data-Warehouse-Design-eBook | 页面 310

BEAM✲ Notation and Short Codes
291
time.
FV Fixed Value attribute. A dimensional attribute that should not change; for
example Date of Birth. FV attributes however can be corrected. When
FV attributes are corrected they behave like CV attributes: the previous
incorrect value is not preserved. 3
PVn Previous Value attribute. A dimensional attribute that records the previous
value of another current value attribute. Also known as a type 3 slowly
changing dimensional attribute.
PVn is always used in conjunction with a matching CVn to relate the previous
value to the current value; for example, Previous Territory PV1 and
Territory CV1.
PV attributes can also be used to hold initial or “as at specific date” values;
for example, Initial Territory PV1 or YE2010 Territory PV1. 6
Event Detail and Fact Column Types
C ODE M EANING /U SAGE
C HAPTERS
MV Multi-Value. Event detail contains multiple values that must be resolved
using a bridge table when converted to a dimensional model.
Fact table FK that points to a multi-value bridge table. 6, 9
ML Multi-Level. Event detail represents various levels in a hierarchy such as
individual employee or teams/branches that must be handled by a multi-level
dimension that contains additional members representing the required levels.
Fact table FK that points to a multi-level dimension and makes use of the
additional levels. 6
DD Degenerate Dimension. Dimensional attribute stored in a fact table. Has no
additional descriptive attributes; therefore, does not join to a physical dimen-
sion table. Typically used for transaction IDs (how details); for example,
Order ID DD. 2, 3, 4, 5
GD
GDn Granular Dimension. A dimension or combination of dimensions that de-
fines the granularity of a fact table.
The numbered version is used when alternative dimension combinations can
define the granularity. For example, Call Reference Number GD1 or
Customer GD2, Call Time GD2 define the granularity of a call detail fact
table. 2, 8
FA (Fully) Additive fact. A fact that produces a correct total when summed
across any combination of its dimensions. For a fact to be additive it must be
expressed in a single unit of measure. Percentages and unit prices are not
additive. 5, 8
SA
SAn Semi-Additive fact. A fact that can be correctly totaled by some dimensions
but not others. Semi-additive facts have at least one non-additive (NA)
dimension. For example, an account balance cannot be summed over time:
its non-additive (NA) dimension. Semi-additive facts are often averaged over
their non-additive dimension.
SA is always used in conjunction with at least one NA dimension foreign key
to relate the semi-additive fact to its non-additive dimensions.
The numbered version is used to relate multiple semi-additive facts in the
same table to their appropriate NA dimensions. For example, Stock Level
SA1 is non-additive across Stock Date Key NA1 whereas Order Count
SA2 is non-additive across Product Key NA2. 8