Oracle
 sql >> Baza danych >  >> RDS >> Oracle

wyrocznia jak obrócić mój stół

jeśli nie chcesz dynamicznej liczby kolumn, musisz zastosować takie podejście i użyć typu do przechowywania kolumn (ponieważ statyczny sql ma stałą liczbę kolumn).

SQL> create table my_table (cy_name varchar2(3), ca_date date, ca_value number);

Table created.

SQL> alter session set nls_date_format='mm/dd/yyyy';

Session altered.

SQL> insert into my_table
  2  select 'CUL' cy_name, to_date('1/4/2013') ca_date, 12 ca_value from dual
  3  union all
  4  select 'CUL' cy_name, to_date('12/31/2012') ca_date, 12 ca_value from dual
  5  union all
  6  select 'CUL' cy_name, to_date('1/3/2013') ca_date, 12 ca_value from dual
  7  union all
  8  select 'CUR' cy_name, to_date('1/6/2013') ca_date, 12 ca_value from dual
  9  union all
 10  select 'CUR' cy_name, to_date('12/6/2013') ca_date, 12 ca_value from dual;

5 rows created.

SQL>
SQL> drop type colTab;

Type dropped.

SQL> drop type colType;

Type dropped.

SQL> create type colType as object (colname varchar2(30), value number);
  2  /

Type created.

SQL> create type colTab as table of colType;
  2  /

Type created.

SQL> alter session set nls_date_format='mm/dd/yyyy';

Session altered.

SQL> col val format a80
SQL> set linesize 90
SQL> select cy_name,
  2         cast(multiset(select ca_date, ca_value
  3                         from my_table d
  4                        where d.cy_name = a.cy_name) as coltab) val
  5    from my_table a
  6  group by cy_name;

CY_ VAL(COLNAME, VALUE)
--- --------------------------------------------------------------------------------
CUR COLTAB(COLTYPE('01/06/2013', 12), COLTYPE('12/06/2013', 12))
CUL COLTAB(COLTYPE('01/04/2013', 12), COLTYPE('12/31/2012', 12), COLTYPE('01/03/2013
    ', 12))

lub wygeneruj najpierw dynamicznie sql..np:

SQL> select 'select CY_NAME, ' from dual
  2  union all
  3  select case when rownum != 1 then ',' end || str
  4    from (select distinct  'MAX(DECODE(CA_DATE, TO_DATE(''' || to_char(ca_date, 'dd-mm-yyyy') || ''', ''DD-MM-YYYY''), CA_VALUE, NULL)) as "' || to_char(ca_date, 'dd-mm-yyyy
') || '"' str
  5            from my_table a)
  6  union all
  7     select 'FROM MY_TABLE
  8  GROUP BY  CY_NAME; ' from dual
  9  ;

'SELECTCY_NAME,'
------------------------------------------------------------------------------------------
select CY_NAME,
MAX(DECODE(CA_DATE, TO_DATE('04-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "04-01-2013"
,MAX(DECODE(CA_DATE, TO_DATE('06-12-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "06-12-2013"
,MAX(DECODE(CA_DATE, TO_DATE('31-12-2012', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "31-12-2012"
,MAX(DECODE(CA_DATE, TO_DATE('03-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "03-01-2013"
,MAX(DECODE(CA_DATE, TO_DATE('06-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "06-01-2013"
FROM MY_TABLE
GROUP BY  CY_NAME;


7 rows selected.

i uruchom go:

SQL> select CY_NAME,
  2  MAX(DECODE(CA_DATE, TO_DATE('04-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "04-01-2013"
  3  ,MAX(DECODE(CA_DATE, TO_DATE('06-12-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "06-12-2013"
  4  ,MAX(DECODE(CA_DATE, TO_DATE('31-12-2012', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "31-12-2012"
  5  ,MAX(DECODE(CA_DATE, TO_DATE('03-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "03-01-2013"
  6  ,MAX(DECODE(CA_DATE, TO_DATE('06-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "06-01-2013"
  7  FROM MY_TABLE
  8  GROUP BY  CY_NAME;

CY_ 04-01-2013 06-12-2013 31-12-2012 03-01-2013 06-01-2013
--- ---------- ---------- ---------- ---------- ----------
CUR                    12                               12
CUL         12                    12         12

AKTUALIZUJ w swojej edycji:jeśli liczba kolumn w Twojej edycji wynosi powiedzmy 7:

SQL> select 'select CY_NAME, ' from dual
  2  union all
  3  select case when rownum != 1 then ',' end || str
  4    from (select 'MAX(DECODE(CA_DATE, TO_DATE('''
  5                 || to_char((d-rownum+1), 'dd-mm-yyyy')
  6                 || ''', ''DD-MM-YYYY''), CA_VALUE, NULL)) as "'
  7                 || to_char((d-rownum+1), 'dd-mm-yyyy') || '"' str
  8            from (select max(ca_date) d from cotton_arv) a
  9          connect by level <= 7 /* change to the number of cols needed */)
 10  union all
 11     select 'FROM V_COTTON_ARV
 12  GROUP BY  CY_NAME; ' from dual
 13  ;

'SELECTCY_NAME,'
------------------------------------------------------------------------------------------------
------------------------
select CY_NAME,
    MAX(DECODE(CA_DATE, TO_DATE('04-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "04-01-2013"
    ,MAX(DECODE(CA_DATE, TO_DATE('03-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "03-01-2013"
    ,MAX(DECODE(CA_DATE, TO_DATE('02-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "02-01-2013"
    ,MAX(DECODE(CA_DATE, TO_DATE('01-01-2013', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "01-01-2013"
    ,MAX(DECODE(CA_DATE, TO_DATE('31-12-2012', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "31-12-2012"
    ,MAX(DECODE(CA_DATE, TO_DATE('30-12-2012', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "30-12-2012"
    ,MAX(DECODE(CA_DATE, TO_DATE('29-12-2012', 'DD-MM-YYYY'), CA_VALUE, NULL)) as "29-12-2012"
    FROM V_COTTON_ARV
    GROUP BY  CY_NAME;


  1. Database
  2.   
  3. Mysql
  4.   
  5. Oracle
  6.   
  7. Sqlserver
  8.   
  9. PostgreSQL
  10.   
  11. Access
  12.   
  13. SQLite
  14.   
  15. MariaDB
  1. Ignorowanie zmiennej podstawienia, gdy warunek nie jest spełniony

  2. Błąd:Ta funkcja nie jest dostępna dla używanej bazy danych

  3. C# Łączenie z formatowaniem Oracle DB DateTime

  4. Jak pobrać informacje o hoście, porcie, identyfikatorze, użytkowniku i haśle w java.sql.Connection?

  5. Co to są widoki zmaterializowane?