Tag Archives: consistency

Deterministic functions, result_cache and operators

Posted on by Posted in deterministic functions, oracle, result_cache 8,768 Page views 0 Comments

In previous posts about caching mechanism of determinstic functions I wrote that cached results are kept only between fetch calls, but there is one exception from this rule: if all function parameters are literals, cached result will not be flushed every fetch call.
Little example with difference:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
SQL> create or replace function f_deterministic(p varchar2)
  2     return varchar2
  3     deterministic
  4  as
  5  begin
  6     dbms_output.put_line(p);
  7     return p;
  8  end;
  9  /
SQL> set arrays 2 feed on;
SQL> set serverout on;
SQL> select
  2     f_deterministic(x) a
  3    ,f_deterministic('literal') b
  4  from (select 'not literal' x
  5        from dual
  6        connect by level<=10
  7       );
 
A                              B
------------------------------ ------------------------------
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
not literal                    literal
 
10 rows selected.
 
not literal
literal
not literal
not literal
not literal
not literal
not literal

As you can see, ‘literal’ was printed once, but ‘not literal’ was printed 6 times, so it was returned from cache 4 times.

Also i want to show the differences in consistency between:
1. Calling a function with determinstic and result_cache;
2. Calling an operator for function with result_cache;
3. Calling an operator for function with deterministic and result_cache;

In this example I will do updates in autonomouse transactions to emulate updates in another session during query execution:

Tables and procedures with updates

Variant 1

Variant 2

Variant 3

Test:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
SQL> set arrays 2;
SQL> select
  2     f1(dummy) variant1
  3    ,o2(dummy) variant2
  4    ,o3(dummy) variant3
  5  from dual
  6  connect by level<=10;
 
  VARIANT1   VARIANT2   VARIANT3
---------- ---------- ----------
         1          1          1
         2          1          1
         2          1          1
         3          1          1
         3          1          1
         4          1          1
         4          1          1
         5          1          1
         5          1          1
         6          1          1
 
10 rows selected.
 
SQL> /
 
  VARIANT1   VARIANT2   VARIANT3
---------- ---------- ----------
         7         11         11
         8         11         11
         8         11         11
         9         11         11
         9         11         11
        10         11         11
        10         11         11
        11         11         11
        11         11         11
        12         11         11
 
10 rows selected.

We can see that function F1 returns same results every 2 execution – it is equal to fetch size(“set arraysize 2”),
operator O2 and O3 return same results for all rows in first query execution, but in the second query executions we can see that they are incremented by 10 – it’s equal to number of rows.
What we can learn from that:
1. The use of the function F1 with result_cache and deterministic reduces function executions, but all function results are inconsistent with query;
2. Operator O2 returns consistent results, but function is always executed because we invalidating result_cache every execution;
3. Operator O3 works as well as operator O2, without considering that function is deterministic.

All tests scripts: tests.zip

Oracle 12c: Inconsistency of Inline “with” functions

Posted on by Posted in 12c, oracle, PL/SQL 2,639 Page views 0 Comments

I was hoping that if inline “with” functions are in the query, so their results will be consistent with it (as operators), but unfortunately such functions returns also inconsistent results as standalone pl/sql functions.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
SQL> create table t as select 1 a from dual;
Table created.
 
SQL> declare
  2    j binary_integer;
  3  begin
  4    dbms_job.submit( j
  5                    ,'begin
  6                        for i in 1..10 loop
  7                          dbms_lock.sleep(1);
  8                          update t set a=a+1;
  9                          commit;
 10                        end loop;
 11                      end;'
 12                   );
 13    commit;
 14  end;
 15  /
 
PL/SQL procedure successfully completed.
 
SQL> with
  2     function f return int is
  3       res int;
  4     begin
  5       dbms_lock.sleep(1);
  6       select a into res from t;
  7       return res;
  8     end;
  9  select
 10     f
 11  from dual
 12  connect by level<=10;
 13  /
 
         F
----------
         1
         1
         1
         2
         3
         4
         5
         6
         7
         8
 
10 rows selected.

Interesting: Jonathan Lewis wrote that inline “deterministic” functions doesn’t use caching mechanism as standalone deterministic functions.