[DERBY-3314] RAND(SEED INTEGER) builtin function always returns the same random value for a given seed.

• Type: Bug
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: 10.3.1.4, 10.3.2.1, 10.4.1.3
• Component/s: SQL
• Labels: derby_triage10_5_2
• Urgency: Normal
• Assignee: Unassigned
• Reporter: Daniel John Debrunner
• Created: 10/Jan/08 18:43
• Updated: 15/Feb/11 21:09

Description

RAND or

{fn RAND(seed)}

exists to match the JDBC specification (section C.1)
RAND(integer) Random floating point for seed integer

Trouble is that Derby creates a new Random() instance for every call leading to the same return value for the same seed. Seems to be useful, the function should return a new random number even when handed the same seed.

Some more specification is probably needed, when does a sequence based upon a seed start?

• first call by any connection
• sequence within a connection
• sequence within a sql context (e.g. procedure call, statement etc.)

Also need to be wary of memory leaks if the engine needs to hold onto Random objects beyond the lifetime of the RAND call.

ij> values rand(3);
1
----------------------
0.731057369148862

1 row selected
ij> values rand(3);
1
----------------------
0.731057369148862

1 row selected
ij> values

{fn rand(3)}

;
1
----------------------
0.731057369148862

1 row selected

Issue Links

• is related to: DERBY-3297 Document SIGN, SQRT, RAND, RANDOM, COSH, COT, SINH and TANH functions.