Standardization means implementing the technical model in a certain specific way. To do so, a common naming convention is in place, which allows identifying objects and columns across all Datavault Builder installations.
Upper- or lower cased naming is based on the used database technology (e.g. postgres is lowercase, Exasol is uppercase).
Naming Conventions¶
Datavault - Main Objects¶
Identifier |
Meaning |
|---|---|
Hub |
|
Satellite |
|
Link |
|
Linksatellite |
|
Transaction Link |
Datavault - Object Subtypes¶
Some of the above objects, mainly satellites, can come in different variants, depending on whether its an effectivity/tracking satellite or a basic satellite.
Identifier |
Meaning |
|---|---|
a |
Altered object (SCD Type 4), containing only changes |
h |
Historized object (SCD Type 2), containing full history |
c |
Current object (SCD Type 1), containing latest state |
o |
Original object (SCD Type 0), containing initian state |
w_trk[df] |
Tracking object, either for full (_w_trkf) or delta (_w_trkd) loads. Historizes implicit deletes from the source |
p |
Prototype |
drv |
Driving View / Table |
Columns¶
Identifier |
Meaning |
|---|---|
bk |
Business Key |
h |
Hash of the business key |
s |
Source of the record |
lt[hcop]? |
Load time |
ls[hcop]? |
Last seen |
ll |
Last load |
_dvb_ |
Column for row number, bk or hash in staging |
el[c]? |
Effective last seen |
v[hcp]? |
Full load validity flag |
raw |
Original column of converted field |