What are Teradata Surrogate Keys?
A surrogate key in Teradata is used to map the natural keys of source systems to a unique key which is usually an integer value.
Usually, one or more natural key columns are mapped to a surrogate key that is worth an INTEGER. Often a consecutive number is generated.
For a surrogate key in Teradata, the ROW_NUMBER() function can be used, or an IDENTITY column. Click here for the article about identity columns:
The advantage of an IDENTITY column is that a new surrogate key is generated by the system as soon as a row is inserted into the key table.
However, it is important to define the IDENTITY column so that no duplicates are created.
For this, the IDENTITY column must be defined with GENERATED ALWAYS and NO CYCLE.
GENERATED always prevents the surrogate key column from being updated, NO CYCLE prevents the reuse of values – If the highest possible number for an integer data type is reached, no new surrogate keys can be generated and an error is reported.
Detailed information about IDENTITY columns is available in table DBC.IdCol
Disaster Reload of Key Tables with IDENTITY columns
If a key table that uses IDENTITY columns has to be reloaded, please pay attention to the following to avoid duplicates:
- The key table needs to be re-created with a new START and MINVALUE numbers to exclude the numbers that were generated before to avoid the same number to be generated again.
- The definition of the identity column needs to be changed from “GENERATED ALWAYS” to “GENERATED BY DEFAULT” to be able to reload the keys that were generated before.
- New surrogate keys will be generated only when NULL values are passed to the identity column.
Here is an example where the highest value of the IDENTITY column is 200,000:
CREATE MULTISET TABLE TheKeys ( NATURAL_KEY VARCHAR(500) CHARACTER SET LATIN NOT CASESPECIFIC NOT NULL, SURROGATE_KEY BIGINT GENERATED ALWAYS AS IDENTITY (START WITH 1 INCREMENT BY 1 MINVALUE -999999999999999 MAXVALUE 999999999999999 NO CYCLE), ) UNIQUE PRIMARY INDEX (NATURAL_KEY);
If we need to reload the key table the following DDL is required:
CREATE MULTISET TABLE TheKeys ( NATURAL_KEY VARCHAR(500) CHARACTER SET LATIN NOT CASESPECIFIC NOT NULL, SURROGATE_KEY BIGINT GENERATED BY DEFAULT AS IDENTITY (START WITH 1 INCREMENT BY 1 MINVALUE 200000 MAXVALUE 999999999999999 NO CYCLE), ) UNIQUE PRIMARY INDEX (NATURAL_KEY);
Why should you use Surrogate Keys?
There are several reasons why it is a good idea to replace the natural keys with Teradata surrogate keys.
Different source systems deliver the same natural key for different information.
You may have more than one source system delivering the same value in the natural key for the same target table but with a different meaning.
This fact requires you to use surrogate keys because keeping the natural key from both sources would not allow you to distinguish the objects.
Different source systems deliver a different natural key for the same information.
But it can also be that the same information comes from different source systems with different natural keys. For example, the customers of a bank may be kept in different source systems. If it is necessary to represent these as a single customer in the data warehouse, surrogate keys are perfectly suited for this.
Unfortunately, many Teradata data warehouse implementations decide against surrogate keys and put the effort of integrating the information into the reports.
This approach will ultimately cause more costs than having a right surrogate key design from the begin (don't forget, such an approach requires to repeat the logic of combining object information into each and every single report).
Exchange of Source System leads to the Delivery of new Natural Keys.
If the natural keys of the source system are used, and you undergo a replacement of the original system you end up in a bad situation.
In this case, let's compare the effort caused by it.
If we have surrogate keys available, only the mapping table (Natural Keys -> Surrogate Keys) has to be adjusted, all surrogate keys have to be assigned once to the new natural keys.
If we only have the natural keys available in our data warehouse, this usually means a complete re-design. The old natural keys must be replaced everywhere by the new natural keys!
In general, you never should tightly couple your data model to the source system structures. An invoice table should be called Invoice and not something like SAP-Invoice. Data Warehousing is not the 1:1 storage of operational data.
At the begin of a project, it is always more agile to avoid surrogate keys and integrate source system tables 1:1.
However, this only shifts costs to the end of the chain. Changes that have to be made afterward are many times more expensive!
Teradata Surrogates Keys and Performance
Surrogate keys, if properly used, may increase your performance as you can replace several natural key columns by an integer column.
Often, natural key columns are character columns. From a performance point of view, this makes no difference for the hashing of the primary index. Where there is a negative effect on performance, however, is with joins, since Teradata must take character sets, etc. into account here.
Depending on the number of columns and data types of the natural key, this can have a negligible impact on performance or a noticeable effect.