Short summary: Queries to ALL_SYNONYMS cause FTS of SYS.OBJ$ which can’t be avoided.
Let’s have a look at a simple query and its execution plan in my test 12.1.0.2 instance. The plan is pretty big, adaptive and uses dynamic sampling. It is composed of two UNION ALL branches. The first branch starts at plan step ID 4, and is not a big deal – just 4 buffer gets and no rows returned right in the start of the branch execution, lines 12-16.
The second part is more interesting. It is composed by the _ALL_SYNONYMS_TREE view, and, as the name suggests, it’s a CONNECT BY on top of a multi-table join. And where are the top query conditions applied? On step 27, after _ALL_SYNONYMS_TREE is fully instantiated. This is the only way to execute such a query, since there’s no good START WITH condition in the CONNECT BY, and the top query conditions logically can’t be pushed into START WITH.
SELECT TABLE_NAME, TABLE_OWNER, DB_LINK FROM ALL_SYNONYMS WHERE OWNER = ‘PUBLIC’ AND SYNONYM_NAME = ‘X’;
Plan hash value: 4035506875
———————————————————————————————————————————-
| Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers |
———————————————————————————————————————————-
| 0 | SELECT STATEMENT | | 1 | | 0 |00:00:00.04 | 398 |
| 1 | VIEW | ALL_SYNONYMS | 1 | 3 | 0 |00:00:00.04 | 398 |
| 2 | SORT UNIQUE | | 1 | 3 | 0 |00:00:00.04 | 398 |
| 3 | UNION-ALL | | 1 | | 0 |00:00:00.04 | 398 |
| * 4 | FILTER | | 1 | | 0 |00:00:00.01 | 4 |
|- * 5 | HASH JOIN | | 1 | 1 | 0 |00:00:00.01 | 4 |
| 6 | NESTED LOOPS | | 1 | 1 | 0 |00:00:00.01 | 4 |
| 7 | NESTED LOOPS | | 1 | 1 | 0 |00:00:00.01 | 4 |
|- 8 | STATISTICS COLLECTOR | | 1 | | 0 |00:00:00.01 | 4 |
|- * 9 | HASH JOIN | | 1 | 1 | 0 |00:00:00.01 | 4 |
| 10 | NESTED LOOPS | | 1 | 1 | 0 |00:00:00.01 | 4 |
|- 11 | STATISTICS COLLECTOR | | 1 | | 0 |00:00:00.01 | 4 |
| 12 | NESTED LOOPS | | 1 | 1 | 0 |00:00:00.01 | 4 |
| 13 | TABLE ACCESS BY INDEX ROWID | USER$ | 1 | 1 | 1 |00:00:00.01 | 2 |
| * 14 | INDEX UNIQUE SCAN | I_USER1 | 1 | 1 | 1 |00:00:00.01 | 1 |
| 15 | TABLE ACCESS BY INDEX ROWID BATCHED| OBJ$ | 1 | 1 | 0 |00:00:00.01 | 2 |
| * 16 | INDEX RANGE SCAN | I_OBJ5 | 1 | 1 | 0 |00:00:00.01 | 2 |
| * 17 | INDEX RANGE SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
|- 18 | INDEX FULL SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 19 | INDEX UNIQUE SCAN | I_SYN1 | 0 | 1 | 0 |00:00:00.01 | 0 |
| 20 | TABLE ACCESS BY INDEX ROWID | SYN$ | 0 | 1 | 0 |00:00:00.01 | 0 |
|- 21 | TABLE ACCESS FULL | SYN$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 22 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| 23 | NESTED LOOPS SEMI | | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 24 | INDEX SKIP SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 25 | INDEX RANGE SCAN | I_OBJ4 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 26 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 27 | VIEW | _ALL_SYNONYMS_TREE | 1 | 2 | 0 |00:00:00.04 | 394 |
| * 28 | CONNECT BY NO FILTERING WITH START-WITH | | 1 | | 0 |00:00:00.04 | 394 |
| * 29 | FILTER | | 1 | | 0 |00:00:00.04 | 394 |
| * 30 | HASH JOIN | | 1 | 89 | 0 |00:00:00.04 | 394 |
| 31 | TABLE ACCESS FULL | USER$ | 1 | 76 | 76 |00:00:00.01 | 6 |
| * 32 | HASH JOIN | | 1 | 89 | 0 |00:00:00.04 | 388 |
| 33 | INDEX FULL SCAN | I_USER2 | 1 | 76 | 76 |00:00:00.01 | 1 |
|- * 34 | HASH JOIN | | 1 | 89 | 0 |00:00:00.03 | 387 |
| 35 | NESTED LOOPS | | 1 | 89 | 0 |00:00:00.03 | 387 |
| 36 | NESTED LOOPS | | 1 | | 0 |00:00:00.03 | 387 |
|- 37 | STATISTICS COLLECTOR | | 1 | | 0 |00:00:00.03 | 387 |
| * 38 | HASH JOIN | | 1 | 89 | 0 |00:00:00.03 | 387 |
| 39 | INDEX FULL SCAN | I_USER2 | 1 | 76 | 76 |00:00:00.01 | 1 |
| * 40 | HASH JOIN | | 1 | 89 | 0 |00:00:00.03 | 386 |
| * 41 | TABLE ACCESS FULL | OBJ$ | 1 | 5182 | 5182 |00:00:00.01 | 349 |
| * 42 | HASH JOIN | | 1 | 5102 | 5178 |00:00:00.01 | 37 |
| 43 | TABLE ACCESS FULL | USER$ | 1 | 76 | 76 |00:00:00.01 | 6 |
| 44 | TABLE ACCESS FULL | SYN$ | 1 | 5102 | 5182 |00:00:00.01 | 31 |
| * 45 | INDEX RANGE SCAN | I_OBJ1 | 0 | | 0 |00:00:00.01 | 0 |
| 46 | TABLE ACCESS BY INDEX ROWID | OBJ$ | 0 | 1 | 0 |00:00:00.01 | 0 |
|- * 47 | TABLE ACCESS FULL | OBJ$ | 0 | 5182 | 0 |00:00:00.01 | 0 |
| * 48 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| 49 | NESTED LOOPS SEMI | | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 50 | INDEX SKIP SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 51 | INDEX RANGE SCAN | I_OBJ4 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 52 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 53 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| 54 | NESTED LOOPS SEMI | | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 55 | INDEX SKIP SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 56 | INDEX RANGE SCAN | I_OBJ4 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 57 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 58 | FILTER | | 0 | | 0 |00:00:00.01 | 0 |
| * 59 | FILTER | | 0 | | 0 |00:00:00.01 | 0 |
| 60 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 61 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 62 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 63 | TABLE ACCESS BY INDEX ROWID | SYN$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 64 | INDEX UNIQUE SCAN | I_SYN1 | 0 | 1 | 0 |00:00:00.01 | 0 |
| 65 | TABLE ACCESS BY INDEX ROWID BATCHED | OBJ$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 66 | INDEX RANGE SCAN | I_OBJ1 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 67 | INDEX RANGE SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 68 | INDEX RANGE SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 69 | FILTER | | 0 | | 0 |00:00:00.01 | 0 |
| * 70 | FILTER | | 0 | | 0 |00:00:00.01 | 0 |
| 71 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 72 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 73 | NESTED LOOPS | | 0 | 1 | 0 |00:00:00.01 | 0 |
| 74 | TABLE ACCESS BY INDEX ROWID | USER$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 75 | INDEX UNIQUE SCAN | I_USER1 | 0 | 1 | 0 |00:00:00.01 | 0 |
| 76 | TABLE ACCESS BY INDEX ROWID BATCHED | OBJ$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 77 | INDEX RANGE SCAN | I_OBJ5 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 78 | INDEX RANGE SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 79 | INDEX RANGE SCAN | I_OBJAUTH1 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 80 | FIXED TABLE FULL | X$KZSRO | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 81 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| 82 | NESTED LOOPS SEMI | | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 83 | INDEX SKIP SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 84 | INDEX RANGE SCAN | I_OBJ4 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 85 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 86 | FIXED TABLE FULL | X$KZSPR | 0 | 23 | 0 |00:00:00.01 | 0 |
| * 87 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
| 88 | NESTED LOOPS SEMI | | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 89 | INDEX SKIP SCAN | I_USER2 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 90 | INDEX RANGE SCAN | I_OBJ4 | 0 | 1 | 0 |00:00:00.01 | 0 |
| * 91 | TABLE ACCESS FULL | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 |
———————————————————————————————————————————-
Predicate Information (identified by operation id):
—————————————————
4 – filter((( IS NULL AND "O"."TYPE#"<>88) OR BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR
(((SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’ AND "U"."TYPE#"<>2) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL) AND IS NOT NULL)))
5 – access("O"."OBJ#"="S"."OBJ#")
9 – access("O"."OWNER#"="U"."USER#")
14 – access("U"."NAME"=’PUBLIC’)
16 – access("O"."SPARE3"="U"."USER#" AND "O"."NAME"=’X’ AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
17 – access("O"."OWNER#"="U"."USER#")
19 – access("O"."OBJ#"="S"."OBJ#")
22 – filter(("TYPE#"=:B1 AND "UE"."USER#"=:B2))
24 – access("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter(("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))))
25 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88 AND "O2"."OWNER#"="U2"."USER#")
26 – filter(("UE"."TYPE#"=:B1 AND "UE"."USER#"=:B2))
27 – filter(("ST"."SYN_OWNER"=’PUBLIC’ AND "ST"."SYN_SYNONYM_NAME"=’X’))
28 – access("S"."BASE_SYN_ID"=PRIOR NULL AND "S"."ORIGIN_CON_ID"=PRIOR NULL)
filter( IS NOT NULL)
29 – filter(((( IS NULL AND "O"."TYPE#"<>88) OR BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR
(((SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’ AND "U"."TYPE#"<>2) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL) AND IS NOT NULL)) AND (( IS NULL
AND "O"."TYPE#"<>88) OR BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR
(((SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’ AND "U"."TYPE#"<>2) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL) AND IS NOT NULL))))
30 – access("O"."SPARE3"="U"."USER#")
32 – access("O"."OWNER#"="U"."USER#")
34 – access("O"."OBJ#"="S"."OBJ#")
38 – access("O"."OWNER#"="U"."USER#")
40 – access("BU"."USER#"="O"."SPARE3" AND "S"."NAME"="O"."NAME")
41 – filter("O"."TYPE#"=5)
42 – access("S"."OWNER"="BU"."NAME")
45 – access("O"."OBJ#"="S"."OBJ#" AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
47 – filter("O"."TYPE#"=5)
48 – filter(("TYPE#"=:B1 AND "UE"."USER#"=:B2))
50 – access("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter(("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))))
51 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88 AND "O2"."OWNER#"="U2"."USER#")
52 – filter(("UE"."TYPE#"=:B1 AND "UE"."USER#"=:B2))
53 – filter(("TYPE#"=:B1 AND "UE"."USER#"=:B2))
55 – access("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter(("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))))
56 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88 AND "O2"."OWNER#"="U2"."USER#")
57 – filter(("UE"."TYPE#"=:B1 AND "UE"."USER#"=:B2))
58 – filter((( IS NOT NULL OR ("S"."NODE" IS NULL AND IS NOT NULL)) AND (( IS NULL AND "O"."TYPE#"<>88) OR
BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR
(((SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’ AND "U"."TYPE#"<>2) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL) AND IS NOT NULL))))
59 – filter(TO_NUMBER(SYS_CONTEXT(‘USERENV’,’CON_ID’)) IS NOT NULL)
64 – access("S"."OBJ#"=:B1)
66 – access("O"."OBJ#"=:B1 AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
67 – access("O"."SPARE3"="U"."USER#")
68 – access("O"."OWNER#"="U"."USER#")
69 – filter((( IS NOT NULL OR "BA"."GRANTOR#"=USERENV(‘SCHEMAID’)) AND (( IS NULL AND "O"."TYPE#"<>88) OR
BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR
(((SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’ AND "U"."TYPE#"<>2) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL) AND IS NOT NULL))))
70 – filter(:B1 IS NULL)
75 – access("BU"."NAME"=:B1)
77 – access("BU"."USER#"="O"."SPARE3" AND "O"."NAME"=:B1)
78 – access("O"."OWNER#"="U"."USER#")
79 – access("BA"."OBJ#"="O"."OBJ#")
80 – filter("KZSROROL"=:B1)
81 – filter(("TYPE#"=:B1 AND "UE"."USER#"=:B2))
83 – access("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter(("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))))
84 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88 AND "O2"."OWNER#"="U2"."USER#")
85 – filter(("UE"."TYPE#"=:B1 AND "UE"."USER#"=:B2))
86 – filter((((-"KZSPRPRV")=(-45) OR (-"KZSPRPRV")=(-47) OR (-"KZSPRPRV")=(-397) OR (-"KZSPRPRV")=(-48) OR
(-"KZSPRPRV")=(-49) OR (-"KZSPRPRV")=(-50)) AND "INST_ID"=USERENV(‘INSTANCE’)))
87 – filter(("TYPE#"=:B1 AND "UE"."USER#"=:B2))
89 – access("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter(("U2"."TYPE#"=2 AND "U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))))
90 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88 AND "O2"."OWNER#"="U2"."USER#")
91 – filter(("UE"."TYPE#"=:B1 AND "UE"."USER#"=:B2))
Note
—–
– dynamic statistics used: dynamic sampling (level=2)
– this is an adaptive plan (rows marked ‘-‘ are inactive)
Because my test database is very small – only 22K objects created – such a query to ALL_SYNONYMS is pretty fast, and it takes just under 400 buffer gets to execute. Let’s look at the same query execution statistics in an EBS 12.2 database. Although the plan is slightly different, its shape is still the same, the hierarchy is still there, and SYS.OBJ$ full scan is present. You can see that OBJ$ is much bigger, more than 600MB. FTS of OBJ$ was pretty fast because of the buffered reads, and all OBJ$‘s blocks were cached already. Occasionally though, Oracle switches to direct path reads while running this query, and it means reading 600MB+ data off disk. In such a case, the execution time can go up; I’ve seen a few cases where it was as high as 30s.
On top of the issue with FTS, there are multiple subqueries which drive buffer gets up to 720K. This is a side effect of filtering _ALL_SYNONYMS_TREE data too late on step 23.
—————————————————————————————————————————————-
| Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers | Reads |
—————————————————————————————————————————————-
| 0 | SELECT STATEMENT | | 1 | | 1 |00:00:04.78 | 720K| 2387 |
| 1 | VIEW | ALL_SYNONYMS | 1 | 3 | 1 |00:00:04.78 | 720K| 2387 |
| 2 | SORT UNIQUE | | 1 | 3 | 1 |00:00:04.78 | 720K| 2387 |
| 3 | UNION-ALL | | 1 | | 1 |00:00:04.78 | 720K| 2387 |
|* 4 | FILTER | | 1 | | 1 |00:00:00.01 | 15 | 2 |
| 5 | NESTED LOOPS | | 1 | 1 | 1 |00:00:00.01 | 12 | 2 |
| 6 | NESTED LOOPS | | 1 | 1 | 1 |00:00:00.01 | 11 | 2 |
| 7 | NESTED LOOPS | | 1 | 1 | 1 |00:00:00.01 | 9 | 1 |
| 8 | NESTED LOOPS | | 1 | 1 | 1 |00:00:00.01 | 7 | 1 |
| 9 | TABLE ACCESS BY INDEX ROWID | USER$ | 1 | 1 | 1 |00:00:00.01 | 3 | 0 |
|* 10 | INDEX UNIQUE SCAN | I_USER1 | 1 | 1 | 1 |00:00:00.01 | 2 | 0 |
| 11 | TABLE ACCESS BY INDEX ROWID BATCHED | OBJ$ | 1 | 1 | 1 |00:00:00.01 | 4 | 1 |
|* 12 | INDEX RANGE SCAN | I_OBJ5 | 1 | 1 | 1 |00:00:00.01 | 3 | 1 |
|* 13 | INDEX RANGE SCAN | I_USER2 | 1 | 1 | 1 |00:00:00.01 | 2 | 0 |
|* 14 | INDEX UNIQUE SCAN | I_SYN1 | 1 | 1 | 1 |00:00:00.01 | 2 | 1 |
| 15 | TABLE ACCESS BY INDEX ROWID | SYN$ | 1 | 1 | 1 |00:00:00.01 | 1 | 0 |
|* 16 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 1 | 1 | 1 |00:00:00.01 | 3 | 0 |
|* 17 | INDEX RANGE SCAN | I_USER_EDITIONING | 1 | 12 | 2 |00:00:00.01 | 2 | 0 |
|* 18 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 0 | 1 | 0 |00:00:00.01 | 0 | 0 |
|* 19 | INDEX RANGE SCAN | I_USER_EDITIONING | 0 | 12 | 0 |00:00:00.01 | 0 | 0 |
| 20 | NESTED LOOPS SEMI | | 0 | 5 | 0 |00:00:00.01 | 0 | 0 |
|* 21 | INDEX RANGE SCAN | I_OBJ4 | 0 | 5 | 0 |00:00:00.01 | 0 | 0 |
|* 22 | INDEX RANGE SCAN | I_USER2 | 0 | 266 | 0 |00:00:00.01 | 0 | 0 |
|* 23 | VIEW | _ALL_SYNONYMS_TREE | 1 | 2 | 0 |00:00:04.78 | 720K| 2385 |
|* 24 | CONNECT BY NO FILTERING WITH START-WITH | | 1 | | 10075 |00:00:04.78 | 720K| 2385 |
|* 25 | FILTER | | 1 | | 10058 |00:00:01.41 | 333K| 0 |
|* 26 | HASH JOIN | | 1 | 70 | 10058 |00:00:00.63 | 112K| 0 |
|* 27 | HASH JOIN | | 1 | 70 | 10058 |00:00:00.61 | 112K| 0 |
| 28 | NESTED LOOPS | | 1 | 70 | 10058 |00:00:00.60 | 112K| 0 |
| 29 | NESTED LOOPS | | 1 | 70 | 10058 |00:00:00.59 | 105K| 0 |
|* 30 | HASH JOIN | | 1 | 70 | 10058 |00:00:00.46 | 87654 | 0 |
|* 31 | HASH JOIN | | 1 | 70 | 10058 |00:00:00.45 | 87634 | 0 |
| 32 | TABLE ACCESS FULL | USER$ | 1 | 3293 | 3293 |00:00:00.01 | 173 | 0 |
|* 33 | HASH JOIN | | 1 | 193K| 198K|00:00:00.42 | 87461 | 0 |
| 34 | TABLE ACCESS FULL | SYN$ | 1 | 174K| 174K|00:00:00.04 | 1086 | 0 |
|* 35 | TABLE ACCESS FULL | OBJ$ | 1 | 174K| 174K|00:00:00.28 | 86375 | 0 |
| 36 | INDEX FULL SCAN | I_USER2 | 1 | 3293 | 3293 |00:00:00.01 | 20 | 0 |
|* 37 | INDEX RANGE SCAN | I_OBJ1 | 10058 | 1 | 10058 |00:00:00.13 | 17611 | 0 |
| 38 | TABLE ACCESS BY INDEX ROWID | OBJ$ | 10058 | 1 | 10058 |00:00:00.01 | 6803 | 0 |
| 39 | INDEX FULL SCAN | I_USER2 | 1 | 3293 | 3293 |00:00:00.01 | 20 | 0 |
| 40 | TABLE ACCESS FULL | USER$ | 1 | 3293 | 3293 |00:00:00.01 | 173 | 0 |
|* 41 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 4 | 1 | 3 |00:00:00.01 | 11 | 0 |
|* 42 | INDEX RANGE SCAN | I_USER_EDITIONING | 4 | 12 | 9 |00:00:00.01 | 8 | 0 |
|* 43 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 3 | 1 | 3 |00:00:00.01 | 9 | 0 |
|* 44 | INDEX RANGE SCAN | I_USER_EDITIONING | 3 | 12 | 9 |00:00:00.01 | 6 | 0 |
| 45 | NESTED LOOPS SEMI | | 9419 | 5 | 9419 |00:00:00.39 | 113K| 0 |
|* 46 | INDEX RANGE SCAN | I_OBJ4 | 9419 | 5 | 81193 |00:00:00.26 | 22421 | 0 |
|* 47 | INDEX RANGE SCAN | I_USER2 | 81193 | 266 | 9419 |00:00:00.11 | 90613 | 0 |
|* 48 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 42 | 1 | 42 |00:00:00.01 | 121 | 0 |
|* 49 | INDEX RANGE SCAN | I_USER_EDITIONING | 42 | 12 | 122 |00:00:00.01 | 79 | 0 |
|* 50 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 41 | 1 | 41 |00:00:00.01 | 118 | 0 |
|* 51 | INDEX RANGE SCAN | I_USER_EDITIONING | 41 | 12 | 120 |00:00:00.01 | 77 | 0 |
| 52 | NESTED LOOPS SEMI | | 10053 | 5 | 10053 |00:00:00.35 | 107K| 0 |
|* 53 | INDEX RANGE SCAN | I_OBJ4 | 10053 | 5 | 69782 |00:00:00.24 | 27630 | 0 |
|* 54 | INDEX RANGE SCAN | I_USER2 | 69782 | 266 | 10053 |00:00:00.09 | 79952 | 0 |
|* 55 | FILTER | | 9429 | | 9424 |00:00:03.29 | 387K| 2385 |
|* 56 | FILTER | | 9429 | | 9424 |00:00:00.37 | 68334 | 211 |
| 57 | NESTED LOOPS | | 9429 | 1 | 9424 |00:00:00.37 | 68334 | 211 |
| 58 | NESTED LOOPS | | 9429 | 1 | 9424 |00:00:00.35 | 58613 | 211 |
| 59 | NESTED LOOPS | | 9429 | 1 | 9424 |00:00:00.33 | 49184 | 211 |
| 60 | TABLE ACCESS BY INDEX ROWID | SYN$ | 9429 | 1 | 9424 |00:00:00.24 | 21904 | 211 |
|* 61 | INDEX UNIQUE SCAN | I_SYN1 | 9429 | 1 | 9424 |00:00:00.22 | 12480 | 211 |
| 62 | TABLE ACCESS BY INDEX ROWID BATCHED | OBJ$ | 9424 | 1 | 9424 |00:00:00.08 | 27280 | 0 |
|* 63 | INDEX RANGE SCAN | I_OBJ1 | 9424 | 1 | 9424 |00:00:00.07 | 17811 | 0 |
|* 64 | INDEX RANGE SCAN | I_USER2 | 9424 | 1 | 9424 |00:00:00.01 | 9429 | 0 |
|* 65 | INDEX RANGE SCAN | I_USER2 | 9424 | 1 | 9424 |00:00:00.01 | 9721 | 0 |
|* 66 | FILTER | | 9390 | | 9288 |00:00:02.70 | 206K| 2174 |
|* 67 | FILTER | | 9390 | | 9561 |00:00:02.46 | 98219 | 2174 |
| 68 | NESTED LOOPS | | 9390 | 6 | 9561 |00:00:02.45 | 98219 | 2174 |
| 69 | NESTED LOOPS | | 9390 | 1 | 9703 |00:00:00.79 | 70185 | 202 |
| 70 | NESTED LOOPS | | 9390 | 1 | 9703 |00:00:00.76 | 59765 | 202 |
| 71 | TABLE ACCESS BY INDEX ROWID | USER$ | 9390 | 1 | 9390 |00:00:00.03 | 22222 | 0 |
|* 72 | INDEX UNIQUE SCAN | I_USER1 | 9390 | 1 | 9390 |00:00:00.02 | 11776 | 0 |
| 73 | TABLE ACCESS BY INDEX ROWID BATCHED| OBJ$ | 9390 | 1 | 9703 |00:00:00.73 | 37543 | 202 |
|* 74 | INDEX RANGE SCAN | I_OBJ5 | 9390 | 1 | 9703 |00:00:00.71 | 27840 | 202 |
|* 75 | INDEX RANGE SCAN | I_USER2 | 9703 | 1 | 9703 |00:00:00.02 | 10420 | 0 |
|* 76 | INDEX RANGE SCAN | I_OBJAUTH1 | 9703 | 7 | 9561 |00:00:01.66 | 28034 | 1972 |
|* 77 | FIXED TABLE FULL | X$KZSRO | 161 | 1 | 2 |00:00:00.01 | 0 | 0 |
|* 78 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 206 | 1 | 153 |00:00:00.01 | 588 | 0 |
|* 79 | INDEX RANGE SCAN | I_USER_EDITIONING | 206 | 12 | 948 |00:00:00.01 | 381 | 0 |
|* 80 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 142 | 1 | 142 |00:00:00.01 | 403 | 0 |
|* 81 | INDEX RANGE SCAN | I_USER_EDITIONING | 142 | 12 | 295 |00:00:00.01 | 261 | 0 |
| 82 | NESTED LOOPS SEMI | | 8531 | 5 | 8531 |00:00:00.20 | 107K| 0 |
|* 83 | INDEX RANGE SCAN | I_OBJ4 | 8531 | 5 | 73472 |00:00:00.06 | 24386 | 0 |
|* 84 | INDEX RANGE SCAN | I_USER2 | 73472 | 266 | 8531 |00:00:00.12 | 82632 | 0 |
|* 85 | FIXED TABLE FULL | X$KZSPR | 1 | 5 | 1 |00:00:00.01 | 0 | 0 |
|* 86 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 3 | 1 | 3 |00:00:00.01 | 9 | 0 |
|* 87 | INDEX RANGE SCAN | I_USER_EDITIONING | 3 | 12 | 9 |00:00:00.01 | 6 | 0 |
|* 88 | TABLE ACCESS BY INDEX ROWID BATCHED | USER_EDITIONING$ | 3 | 1 | 3 |00:00:00.01 | 9 | 0 |
|* 89 | INDEX RANGE SCAN | I_USER_EDITIONING | 3 | 12 | 9 |00:00:00.01 | 6 | 0 |
| 90 | NESTED LOOPS SEMI | | 9419 | 5 | 9419 |00:00:00.17 | 113K| 0 |
|* 91 | INDEX RANGE SCAN | I_OBJ4 | 9419 | 5 | 81193 |00:00:00.04 | 22421 | 0 |
|* 92 | INDEX RANGE SCAN | I_USER2 | 81193 | 266 | 9419 |00:00:00.11 | 90613 | 0 |
—————————————————————————————————————————————-
Predicate Information (identified by operation id):
—————————————————
4 – filter((( IS NULL AND "O"."TYPE#"<>88) OR BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR ( IS NOT
NULL AND (("U"."TYPE#"<>2 AND SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL))))
10 – access("U"."NAME"=’PUBLIC’)
12 – access("O"."SPARE3"="U"."USER#" AND "O"."NAME"=’X’ AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
13 – access("O"."OWNER#"="U"."USER#")
14 – access("O"."OBJ#"="S"."OBJ#")
16 – filter("TYPE#"=:B1)
17 – access("UE"."USER#"=:B1)
18 – filter("UE"."TYPE#"=:B1)
19 – access("UE"."USER#"=:B1)
21 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88)
22 – access("O2"."OWNER#"="U2"."USER#" AND "U2"."TYPE#"=2 AND
"U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter("U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
23 – filter(("ST"."SYN_OWNER"=’PUBLIC’ AND "ST"."SYN_SYNONYM_NAME"=’X’))
24 – access("S"."BASE_SYN_ID"=PRIOR NULL AND "S"."ORIGIN_CON_ID"=PRIOR NULL)
filter( IS NOT NULL)
25 – filter(((( IS NULL AND "O"."TYPE#"<>88) OR BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR ( IS NOT
NULL AND (("U"."TYPE#"<>2 AND SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL))) AND (( IS NULL AND "O"."TYPE#"<>88) OR
BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR ( IS NOT NULL AND (("U"."TYPE#"<>2 AND
SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL)))))
26 – access("O"."SPARE3"="U"."USER#")
27 – access("O"."OWNER#"="U"."USER#")
30 – access("O"."OWNER#"="U"."USER#")
31 – access("S"."OWNER"="BU"."NAME" AND "BU"."USER#"="O"."SPARE3")
33 – access("S"."NAME"="O"."NAME")
35 – filter("O"."TYPE#"=5)
37 – access("O"."OBJ#"="S"."OBJ#" AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
41 – filter("TYPE#"=:B1)
42 – access("UE"."USER#"=:B1)
43 – filter("UE"."TYPE#"=:B1)
44 – access("UE"."USER#"=:B1)
46 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88)
47 – access("O2"."OWNER#"="U2"."USER#" AND "U2"."TYPE#"=2 AND
"U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter("U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
48 – filter("TYPE#"=:B1)
49 – access("UE"."USER#"=:B1)
50 – filter("UE"."TYPE#"=:B1)
51 – access("UE"."USER#"=:B1)
53 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88)
54 – access("O2"."OWNER#"="U2"."USER#" AND "U2"."TYPE#"=2 AND
"U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter("U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
55 – filter((( IS NOT NULL OR ("S"."NODE" IS NULL AND IS NOT NULL)) AND (( IS NULL AND "O"."TYPE#"<>88) OR
BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR ( IS NOT NULL AND (("U"."TYPE#"<>2 AND
SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL)))))
56 – filter(TO_NUMBER(SYS_CONTEXT(‘USERENV’,’CON_ID’)) IS NOT NULL)
61 – access("S"."OBJ#"=:B1)
63 – access("O"."OBJ#"=:B1 AND "O"."TYPE#"=5)
filter("O"."TYPE#"=5)
64 – access("O"."SPARE3"="U"."USER#")
65 – access("O"."OWNER#"="U"."USER#")
66 – filter((( IS NOT NULL OR "BA"."GRANTOR#"=USERENV(‘SCHEMAID’)) AND (( IS NULL AND "O"."TYPE#"<>88) OR
BITAND("O"."FLAGS",1048576)=1048576 OR BITAND("U"."SPARE1",16)=0 OR ( IS NOT NULL AND (("U"."TYPE#"<>2 AND
SYS_CONTEXT(‘userenv’,’current_edition_name’)=’ORA$BASE’) OR ("U"."TYPE#"=2 AND
"U"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’))) OR IS NOT NULL)))))
67 – filter(:B1 IS NULL)
72 – access("BU"."NAME"=:B1)
74 – access("BU"."USER#"="O"."SPARE3" AND "O"."NAME"=:B1)
75 – access("O"."OWNER#"="U"."USER#")
76 – access("BA"."OBJ#"="O"."OBJ#")
77 – filter("KZSROROL"=:B1)
78 – filter("TYPE#"=:B1)
79 – access("UE"."USER#"=:B1)
80 – filter("UE"."TYPE#"=:B1)
81 – access("UE"."USER#"=:B1)
83 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88)
84 – access("O2"."OWNER#"="U2"."USER#" AND "U2"."TYPE#"=2 AND
"U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter("U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
85 – filter((((-"KZSPRPRV")=(-45) OR (-"KZSPRPRV")=(-47) OR (-"KZSPRPRV")=(-397) OR (-"KZSPRPRV")=(-48) OR
(-"KZSPRPRV")=(-49) OR (-"KZSPRPRV")=(-50)) AND "INST_ID"=USERENV(‘INSTANCE’)))
86 – filter("TYPE#"=:B1)
87 – access("UE"."USER#"=:B1)
88 – filter("UE"."TYPE#"=:B1)
89 – access("UE"."USER#"=:B1)
91 – access("O2"."DATAOBJ#"=:B1 AND "O2"."TYPE#"=88)
92 – access("O2"."OWNER#"="U2"."USER#" AND "U2"."TYPE#"=2 AND
"U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
filter("U2"."SPARE2"=TO_NUMBER(SYS_CONTEXT(‘userenv’,’current_edition_id’)))
Why CONNECT BY is in this view? It wasn’t always this way. Some time ago, there was no hierarchy inside. It appeared as a fix for bug 3369744 in 10.2.0.1. The fix was supposed to help in situations where ALL_SYNONYMS didn’t report synonyms for synonyms. Hence, Oracle decided to add a hierarchy inside this view. In my opinion, that was a wrong decision. The number of cases where you have multiple layers of synonyms is relatively small. So it would be much better to close this issue as not a bug, and suggest using existing ALL_SYNONYMS to build a tree in the client code. This way a query would run instantly with no FTS because of a perfect START WITH condition.
Oracle JDBC driver as part of its implementation runs a few queries to ALL_SYNONYMS. This is how I found it. The driver uses hierarchical queries on top of the view because it was implemented a long time ago when there was no tree inside ALL_SYNONYMS. I think that was the proper way to deal with the “synonyms for synonyms” case. From what I understand, JDBC driver can run such queries when user code tries to utilize Abstract Data Types/Collections. It caches results per Connection so that the number of times such queries are run is relatively small. If you reuse your Connections and do not re-initialize them too often, of course.