Data Type Evaluation - Teradata Vantage

Teradata® VantageCloud Lake
Deployment
VantageCloud
Edition
Lake
Product
Teradata Vantage
Published
January 2023
Language
English (United States)
Last Update
2024-04-03
dita:mapPath
phg1621910019905.ditamap
dita:ditavalPath
pny1626732985837.ditaval
dita:id
phg1621910019905

Different data types define equality and inequality differently. The following table explains the foundations for how the data types are compared.

Data Type How Evaluated
Numeric Algebraically, with negatives considered to be smaller irrespective of their absolute value.
Byte Bit-by-bit from left to right. A 0 bit is less than a 1 bit.
  • If every pair comparison is equal, then the two byte string are equal.
  • If any pairwise comparison is not equal, then that comparison determines the result.
  • If two byte strings of different lengths are compared, then the shorter string is padded to the right with binary zeros to make the lengths equal before making the comparison.
Character Character-by-character from left to right. Exact comparisons depend on the collation sequence assigned and whether the comparison is case specific or case blind.
The available collations are:
  • ASCII
  • EBCDIC
  • MULTINATIONAL
  • CHARSET_COLL
  • JIS_COLL

If every pairwise comparison is equal, then the two character strings are equal.

If any pairwise comparison is not equal, then that comparison determines the result.

See Comparison of Character Strings of Unequal Length.
DateTime Chronologically.

For information on how Time Zone affects Time comparison, see Time Zone Sort Order.
Interval According to sign and magnitude.
Period Assuming p1 and p2 are Period expressions or derived periods, the evaluation of a Period comparison predicate uses the following logic:

IF BEGIN(p1) = BEGIN(p2) is TRUE, return END(p1) operator END(p2)

ELSE return (BEGIN(p1) operator BEGIN(p2))

For details on BEGIN and END, see Period Functions and Operators.
UDT According to the ordering definition of the UDT.

Vantage generates ordering functionality for distinct UDTs where the source types are not LOBs.

Structured/Distinct/Variant UDTs are only supported on the Block File System on the primary cluster. They are not available for the Object File System.