Administrace komentářů

Jste na stránce určené pro řešení chyb a problémů týkajících se diskusí a komentářů. Můžete zde našim administrátorům reportovat špatně zařazenou či duplicitní diskusi, vulgární či osočující příspěvek a podobně. Děkujeme vám za vaši pomoc, více očí více vidí, společně můžeme udržet vysokou kvalitu AbcLinuxu.cz.

Příspěvek

8.9.2009 20:45 default | skóre: 22 | Madrid
Rozbalit Rozbalit vše Re: SELECT vicero výrazů

Přílohy:

with-example.sql (0 bytů)
with-example.out (0 bytů)

Pointa WITH (factoring klauzule) je hlavně v tom, že materializuje. Bez indexů. A z toho vyplývá i využití. :-D Třeba Oracle umí udělat in-line view, ale někdy ne, což je zákeřné. :-)

Následující příklad ukazuje in-line view:

CREATE TABLE my_all_tables (
    table_id NUMBER(10) NOT NULL,
    table_name VARCHAR2(30) NOT NULL)
/

ALTER TABLE my_all_tables ADD CONSTRAINT pk_mat_tid PRIMARY KEY (table_id)
/

CREATE TABLE my_all_schemas (
    schema_name VARCHAR2(30) NOT NULL,
    table_id NUMBER(10) NOT NULL)
/

INSERT ALL
    INTO my_all_tables (table_id, table_name)
        VALUES (table_id1, table_name1)
    INTO my_all_schemas (schema_name, table_id)
        VALUES (table_schema1, table_id1)
SELECT
    ROW_NUMBER() OVER (ORDER BY att.table_name ASC) AS table_id1,
    att.owner AS table_schema1,
    att.table_name AS table_name1
FROM
    all_tables att
/

COMMIT
/

ALTER TABLE my_all_schemas ADD CONSTRAINT fk_mas_mat FOREIGN KEY (table_id) REFERENCING my_all_tables (table_id)
/

CREATE INDEX idx_mat_tn ON my_all_tables (table_name)
/

EXPLAIN PLAN FOR SELECT
    '"' || mas.schema_name || '"."' || mat.table_name || '"' AS full_name
FROM
    my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id)
WHERE
    mat.table_name LIKE 'C%'
/

SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY)
/

EXPLAIN PLAN FOR SELECT
    '"' || dta.schema_name || '"."' || dta.table_name || '"' AS full_name
FROM
    (SELECT
        mas.schema_name,
        mat.table_name
    FROM
        my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id)) dta
WHERE
    dta.table_name LIKE 'C%'
/

SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY)
/

EXPLAIN PLAN FOR WITH dta AS (SELECT
        mas.schema_name,
        mat.table_name
    FROM
        my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id))
SELECT
    '"' || dta.schema_name || '"."' || dta.table_name || '"' AS full_name
FROM
    dta
WHERE
    dta.table_name LIKE 'C%'
/

SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY)
/

DROP TABLE my_all_schemas PURGE
/

DROP TABLE my_all_tables PURGE
/

A výstup:

CREATE TABLE succeeded.
 ALTER TABLE my_all_tables succeeded.
CREATE TABLE succeeded.
5254 rows inserted
COMMIT succeeded.
 ALTER TABLE my_all_schemas succeeded.
CREATE TABLE succeeded.
 ALTER TABLE my_all_tables succeeded.
CREATE TABLE succeeded.
5254 rows inserted
COMMIT succeeded.
 ALTER TABLE my_all_schemas succeeded.
CREATE INDEX succeeded.
EXPLAIN PLAN succeeded.
PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------------
Plan hash value: 1386169526

-----------------------------------------------------------------------------------------------
| Id  | Operation                    | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |                |   171 | 10260 |    11  (10)| 00:00:01 |
|*  1 |  HASH JOIN                   |                |   171 | 10260 |    11  (10)| 00:00:01 |
|   2 |   TABLE ACCESS BY INDEX ROWID| MY_ALL_TABLES  |   171 |  5130 |     5   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN          | IDX_MAT_TN     |   171 |       |     3   (0)| 00:00:01 |
|   4 |   TABLE ACCESS FULL          | MY_ALL_SCHEMAS |  2627 | 78810 |     5   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
   3 - access("MAT"."TABLE_NAME" LIKE 'C%')
       filter("MAT"."TABLE_NAME" LIKE 'C%')

Note
-----
   - dynamic sampling used for this statement

22 rows selected

EXPLAIN PLAN succeeded.
PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------------
Plan hash value: 1386169526

-----------------------------------------------------------------------------------------------
| Id  | Operation                    | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |                |   171 | 10260 |    11  (10)| 00:00:01 |
|*  1 |  HASH JOIN                   |                |   171 | 10260 |    11  (10)| 00:00:01 |
|   2 |   TABLE ACCESS BY INDEX ROWID| MY_ALL_TABLES  |   171 |  5130 |     5   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN          | IDX_MAT_TN     |   171 |       |     3   (0)| 00:00:01 |
|   4 |   TABLE ACCESS FULL          | MY_ALL_SCHEMAS |  2627 | 78810 |     5   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
   3 - access("MAT"."TABLE_NAME" LIKE 'C%')
       filter("MAT"."TABLE_NAME" LIKE 'C%')

Note
-----
   - dynamic sampling used for this statement

22 rows selected

EXPLAIN PLAN succeeded.
PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------------
Plan hash value: 1386169526

-----------------------------------------------------------------------------------------------
| Id  | Operation                    | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |                |   171 | 10260 |    11  (10)| 00:00:01 |
|*  1 |  HASH JOIN                   |                |   171 | 10260 |    11  (10)| 00:00:01 |
|   2 |   TABLE ACCESS BY INDEX ROWID| MY_ALL_TABLES  |   171 |  5130 |     5   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN          | IDX_MAT_TN     |   171 |       |     3   (0)| 00:00:01 |
|   4 |   TABLE ACCESS FULL          | MY_ALL_SCHEMAS |  2627 | 78810 |     5   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
   3 - access("MAT"."TABLE_NAME" LIKE 'C%')
       filter("MAT"."TABLE_NAME" LIKE 'C%')

Note
-----
   - dynamic sampling used for this statement

22 rows selected

 DROP TABLE my_all_schemas succeeded.
 DROP TABLE my_all_tables succeeded.

Ale jakmile uděláš něco složitějšího:

WITH dta AS (SELECT
        mas.schema_name,
        mat.table_name,
        LENGTH(mat.table_name) AS table_name_len
    FROM
        my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id))
SELECT
    '"' || dta.schema_name || '"."' || dta.table_name || '" (' || gen.val || ')' AS full_name
FROM
    dta INNER JOIN (SELECT
            gen.val
        FROM
            (SELECT
                LEVEL AS val
            FROM
                dual
            CONNECT BY
                LEVEL <= (SELECT MAX(dta.table_name_len) FROM dta)) gen) gen
        ON ((dta.table_name_len <= 1 AND gen.val = 1) OR (dta.table_name_len > 1 AND gen.val <= dta.table_name_len))
WHERE
    dta.table_name LIKE 'C%'
/

pak LOAD AS SELECT a čau:

EXPLAIN PLAN succeeded.
PLAN_TABLE_OUTPUT
-------------------------------------------------------------------------------------------------------------
Plan hash value: 330330509

-------------------------------------------------------------------------------------------------------------
| Id  | Operation                       | Name                      | Rows  | Bytes | Cost (%CPU)| Time     |
-------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                |                           |     8 |   480 |    21   (5)| 00:00:01 |
|   1 |  TEMP TABLE TRANSFORMATION      |                           |       |       |            |          |
|   2 |   LOAD AS SELECT                |                           |       |       |            |          |
|*  3 |    HASH JOIN                    |                           |  2627 |   153K|    11  (10)| 00:00:01 |
|   4 |     TABLE ACCESS FULL           | MY_ALL_SCHEMAS            |  2627 | 78810 |     5   (0)| 00:00:01 |
|   5 |     TABLE ACCESS FULL           | MY_ALL_TABLES             |  2627 | 78810 |     5   (0)| 00:00:01 |
|   6 |   NESTED LOOPS                  |                           |     8 |   480 |    10   (0)| 00:00:01 |
|   7 |    VIEW                         |                           |     1 |    13 |     2   (0)| 00:00:01 |
|*  8 |     CONNECT BY WITHOUT FILTERING|                           |       |       |            |          |
|   9 |      FAST DUAL                  |                           |     1 |       |     2   (0)| 00:00:01 |
|  10 |      SORT AGGREGATE             |                           |     1 |    13 |            |          |
|  11 |       VIEW                      |                           |  2627 | 34151 |     8   (0)| 00:00:01 |
|  12 |        TABLE ACCESS FULL        | SYS_TEMP_0FD9D6608_51DED0 |  2627 |   120K|     8   (0)| 00:00:01 |
|* 13 |    VIEW                         |                           |     8 |   376 |     8   (0)| 00:00:01 |
|  14 |     TABLE ACCESS FULL           | SYS_TEMP_0FD9D6608_51DED0 |  2627 |   120K|     8   (0)| 00:00:01 |
-------------------------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   3 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
   8 - filter(LEVEL<= (SELECT MAX("DTA"."TABLE_NAME_LEN") FROM  (SELECT /*+ CACHE_TEMP_TABLE ("T1")
              */ "C0" "SCHEMA_NAME","C1" "TABLE_NAME","C2" "TABLE_NAME_LEN" FROM "SYS"."SYS_TEMP_0FD9D6608_51DED0"
              "T1") "DTA"))
  13 - filter("DTA"."TABLE_NAME" LIKE 'C%' AND ("DTA"."TABLE_NAME_LEN"<=1 AND "GEN"."VAL"=1 OR
              "DTA"."TABLE_NAME_LEN">1 AND "GEN"."VAL"<="DTA"."TABLE_NAME_LEN"))

Note
-----
   - dynamic sampling used for this statement

35 rows selected

Zatímco, když si ty dotazy rozkopíruješ:

SELECT
    '"' || dta.schema_name || '"."' || dta.table_name || '" (' || gen.val || ')' AS full_name
FROM
    (SELECT
        mas.schema_name,
        mat.table_name,
        LENGTH(mat.table_name) AS table_name_len
    FROM
        my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id)) dta
    INNER JOIN (SELECT
            gen.val
        FROM
            (SELECT
                LEVEL AS val
            FROM
                dual
            CONNECT BY
                LEVEL <= (SELECT
                        MAX(dta.table_name_len)
                    FROM
                        (SELECT
                            mas.schema_name,
                            mat.table_name,
                            LENGTH(mat.table_name) AS table_name_len
                        FROM
                            my_all_schemas mas INNER JOIN my_all_tables mat ON (mas.table_id = mat.table_id)) dta)) gen) gen
        ON ((dta.table_name_len <= 1 AND gen.val = 1) OR (dta.table_name_len > 1 AND gen.val <= dta.table_name_len))
WHERE
    dta.table_name LIKE 'C%'
/

dostaneš:

EXPLAIN PLAN succeeded.
PLAN_TABLE_OUTPUT
---------------------------------------------------------------------------------------------------
Plan hash value: 2113348083

---------------------------------------------------------------------------------------------------
| Id  | Operation                        | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                 |                |     1 |    73 |    12   (9)| 00:00:01 |
|*  1 |  HASH JOIN                       |                |     1 |    73 |    12   (9)| 00:00:01 |
|   2 |   NESTED LOOPS                   |                |       |       |            |          |
|   3 |    NESTED LOOPS                  |                |     1 |    43 |     6   (0)| 00:00:01 |
|   4 |     VIEW                         |                |     1 |    13 |     2   (0)| 00:00:01 |
|*  5 |      CONNECT BY WITHOUT FILTERING|                |       |       |            |          |
|   6 |       FAST DUAL                  |                |     1 |       |     2   (0)| 00:00:01 |
|   7 |       SORT AGGREGATE             |                |     1 |    43 |            |          |
|*  8 |        HASH JOIN                 |                |  2627 |   110K|    11  (10)| 00:00:01 |
|   9 |         TABLE ACCESS FULL        | MY_ALL_SCHEMAS |  2627 | 34151 |     5   (0)| 00:00:01 |
|  10 |         TABLE ACCESS FULL        | MY_ALL_TABLES  |  2627 | 78810 |     5   (0)| 00:00:01 |
|* 11 |     INDEX RANGE SCAN             | IDX_MAT_TN     |   171 |       |     2   (0)| 00:00:01 |
|* 12 |    TABLE ACCESS BY INDEX ROWID   | MY_ALL_TABLES  |     1 |    30 |     4   (0)| 00:00:01 |
|  13 |   TABLE ACCESS FULL              | MY_ALL_SCHEMAS |  2627 | 78810 |     5   (0)| 00:00:01 |
---------------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
   5 - filter(LEVEL<= (SELECT MAX(LENGTH("MAT"."TABLE_NAME")) FROM "MY_ALL_TABLES"
              "MAT","MY_ALL_SCHEMAS" "MAS" WHERE "MAS"."TABLE_ID"="MAT"."TABLE_ID"))
   8 - access("MAS"."TABLE_ID"="MAT"."TABLE_ID")
  11 - access("MAT"."TABLE_NAME" LIKE 'C%')
       filter("MAT"."TABLE_NAME" LIKE 'C%')
  12 - filter("GEN"."VAL"=1 AND LENGTH("MAT"."TABLE_NAME")<=1 OR
              LENGTH("MAT"."TABLE_NAME")>1 AND "GEN"."VAL"<=LENGTH("MAT"."TABLE_NAME"))

Note
-----
   - dynamic sampling used for this statement

36 rows selected

Ano, příklad není ideální, ale nic lepšího mě teď narychlo nenapadá.

Oracle 11gR1

V tomto formuláři můžete formulovat svou stížnost ohledně příspěvku. Nejprve vyberte typ akce, kterou navrhujete provést s diskusí či příspěvkem. Potom do textového pole napište důvody, proč by měli admini provést vaši žádost, problém nemusí být patrný na první pohled. Odkaz na příspěvek bude přidán automaticky.

Offtopic diskuse použijte pro diskusi mimo záběr našeho portálu (včelařství, windows), která by měla být smazána.
Duplicitní diskuse je určena pro případ, kdy uživatel odeslal svůj dotaz několikrát.
Návrh na cenzuru použijte, pokud komentář obsahuje urážky, vulgarismy nebo porušuje zákony.
Návrh na smazání komentáře je velmi neobvyklá akce používaná obvykle při nepovolené reklamě či spamu, výkřicích psychicky nemocných jedinců (ještírci) nebo pro smazání duplikátních komentářů.
Oprava formátování je žádost pro úpravu formy příspěvku, například pokud někdo píše bez odstavců nebo použije PRE pro celý komentář. Žádosti o změnu obsahu budou zamítnuty.

Vaše jméno
Váš email
Typ požadavku

Slovní popis