È necessario trovare il tempo di elaborazione medio tra tutti i record di timestamp in Oracle SQL

L'esecuzione di intervalli di tempo, e in particolare di aggregazioni (somme, medie) su molti intervalli o gruppi di intervalli, non è semplice in Oracle. La funzione AVG non funziona su intervalli di timestamp, prevede numeri. Quindi, dovremo creare il nostro oggetto di aggregazione e la nostra funzione che lo farà.

In primo luogo, le specifiche dell'oggetto :

CREATE OR REPLACE TYPE AvgInterval 
AS OBJECT (
runningSum INTERVAL DAY(9) TO SECOND(9),
runningCnt number,

STATIC FUNCTION ODCIAggregateInitialize
  ( actx IN OUT AvgInterval
  ) RETURN NUMBER,

MEMBER FUNCTION ODCIAggregateIterate
  ( self  IN OUT AvgInterval,
    val   IN       DSINTERVAL_UNCONSTRAINED
  ) RETURN NUMBER,

MEMBER FUNCTION ODCIAggregateTerminate
  ( self             IN   AvgInterval,
    returnValue  OUT DSINTERVAL_UNCONSTRAINED,
    flags           IN   NUMBER
  ) RETURN NUMBER,

MEMBER FUNCTION ODCIAggregateMerge
  (self  IN OUT AvgInterval,
   ctx2 IN      AvgInterval
  ) RETURN NUMBER
);

E il corpo dell'oggetto :

CREATE OR REPLACE TYPE BODY AvgInterval AS
STATIC FUNCTION ODCIAggregateInitialize
  ( actx IN OUT AvgInterval
  ) RETURN NUMBER IS 
  BEGIN
    IF actx IS NULL THEN
      actx := AvgInterval (INTERVAL '0 0:0:0.0' DAY TO SECOND, 0);
    ELSE
      actx.runningSum := INTERVAL '0 0:0:0.0' DAY TO SECOND;
      actx.runningCnt := 0;
    END IF;
    RETURN ODCIConst.Success;
  END;

MEMBER FUNCTION ODCIAggregateIterate
  ( self  IN OUT AvgInterval,
    val   IN     DSINTERVAL_UNCONSTRAINED
  ) RETURN NUMBER IS
  BEGIN
    self.runningSum := self.runningSum + val;
    self.runningCnt := self.runningCnt + 1;
    RETURN ODCIConst.Success;
  END;

MEMBER FUNCTION ODCIAggregateTerminate
  ( self        IN  AvgInterval,
    ReturnValue OUT DSINTERVAL_UNCONSTRAINED,
    flags       IN  NUMBER
  ) RETURN NUMBER IS
  BEGIN
    if (runningCnt <> 0) then
        returnValue := (self.runningSum/runningCnt);
    else
        returnValue := self.runningSum;
    end if;
    RETURN ODCIConst.Success;
  END;

MEMBER FUNCTION ODCIAggregateMerge
  (self IN OUT AvgInterval,
   ctx2 IN     AvgInterval
  ) RETURN NUMBER IS
  BEGIN
    self.runningSum := self.runningSum + ctx2.runningSum;
    self.runningCnt := self.runningCnt + ctx2.runningCnt;
    RETURN ODCIConst.Success;
  END;

END;

Infine, la funzione che utilizza l'oggetto sopra :

CREATE OR REPLACE FUNCTION avg_interval( x DSINTERVAL_UNCONSTRAINED) 
RETURN DSINTERVAL_UNCONSTRAINED  PARALLEL_ENABLE
AGGREGATE USING AvgInterval;

Ora possiamo usarlo così :

with x as (
    select systimestamp - 1/24 as created_date, systimestamp as modified_date from dual
    union
    select systimestamp - 2/24 as created_date, systimestamp as modified_date from dual
    union
    select systimestamp - 3/24 as created_date, systimestamp as modified_date from dual
)
select avg_interval(modified_date - created_date)
from x;

Uscita:

+00 02:00:00.562669

Possiamo anche aggregare su gruppi con questo :

with x as (
    select 'FL' as state, to_dsinterval('0 00:56:30') as duration from dual
    union
    select 'FL' as state, to_dsinterval('0 02:08:40') as duration from dual
    union
    select 'GA' as state, to_dsinterval('0 01:01:00') as duration from dual
)
select state, avg_interval(duration)
from x
group by state;

Uscita:

FL  +00 01:32:35.000000
GA  +00 01:01:00.000000