Advertisements

Differences between Unicode and EBCDIC sorting sequences

Tags

, , , , , , ,

Recently I came to know that DB2 sorts the data that it returns based on how the tables are encoded when it is created.

 For example, the DB2 catalog is stored in Unicode, any queries that you issue against Unicode tables in the catalog use the Unicode sorting sequence

EBCDICUnicode and ASCII
CharactersHexadecimal valueCharactersHexadecimal value
spaceX’40’spaceX’20’
lowercase charactersX’81’ – X’89’
X’91’ – X’99’
X’A1′ – X’A9′
numeralsX’30’ – X’39’
uppercase charactersX’C1′ – X’C9′
X’D1′ – X’D9′
X’E1′ – X’E9′
uppercase charactersX’40’ – X’4F’
X’50’ – X5A’
numeralsX’F0′ – X’F9′lowercase charactersX’61’ – X’6F’
X’70’ – X7A’

Equal predicates are not affected by the different sorting sequences.

Examples

For the following examples, assume that a table called MYTABLES has a NAME column that is type VARCHAR(128). This column contains the following values: TEST1, TEST2, TEST3, TESTA, TESTB, and TESTC.Example query with ORDER BYSuppose that you issue the following SQL query:

SELECT NAME FROM MYTABLES
ORDER BY NAME

If MYTABLES is encoded in Unicode, DB2 returns the following result:

TEST1
TEST2
TEST3
TESTA
TESTB
TESTC

If MYTABLES is encoded in EBCDIC, DB2 returns the following result:

TESTA
TESTB
TESTC
TEST1
TEST2
TEST3

Example of query with ORDER BY and BETWEEN predicateAssume that you issue the following SQL query:

SELECT * FROM MYTABLES
WHERE NAME BETWEEN 'TEST2' AND 'TESTB'
ORDER BY NAME

If MYTABLES is encoded in Unicode, DB2 returns the following result:

TEST3
TESTA

If MYTABLES is encoded in EBCDIC, DB2 returns the following result:

TESTC
TEST1

Note: Information is pulled from https://www.ibm.com/support/knowledgecenter/en/SSEPEK_10.0.0/char/src/tpc/db2z_diffuniebcdicsortseq.html

Advertisements