forall分批提交oracle,ORACLE中 用Forall与bulk collect快速复制表数据

本文中介绍的几种写法分别是从代码的简易性,FORALL和bulk collect的使用,以及分批插入这三方面考虑得出的，大家可以根据自己的需要灵活选择。

三种不同的写法:

1.使用了BULK COLLECT,没有使用FORALL, 一次性插入,分批COMMIT,这种方法比较适用于10万以下条数据的表;

create or replace procedure cp_data2 as

type TYPE_EMPLOYEES is table of EMPLOYEES%rowtype;

V_EMPLOYEES TYPE_EMPLOYEES;

v_table varchar2(30);

v_sql varchar2(300);

v_rows number:=5000;

begin

execute immediate 'alter session set nls_date_format=''yyyy/mm/dd''';

v_table := 'employee_cp';

v_sql := 'insert /*+ APPEND*/ into ' || v_table ||

' (EMPLOYEE_ID,

FIRST_NAME,

LAST_NAME,

EMAIL,

PHONE_NUMBER,

HIRE_DATE,

JOB_ID,

SALARY,

COMMISSION_PCT,

MANAGER_ID,

DEPARTMENT_ID,

BIRTHDAY)

values (:1, :2,:3,:4,:5,:6, :7, :8,:9,:10, :11,:12)';

select * bulk collect into V_EMPLOYEES from employees; --dest table

for i in 1 .. V_EMPLOYEES.count loop

execute immediate v_sql

using V_EMPLOYEES(i).EMPLOYEE_ID, V_EMPLOYEES(i).FIRST_NAME, V_EMPLOYEES(i).LAST_NAME, V_EMPLOYEES(i).EMAIL, V_EMPLOYEES(i).PHONE_NUMBER, V_EMPLOYEES(i).HIRE_DATE, V_EMPLOYEES(i).JOB_ID, V_EMPLOYEES(i).SALARY, V_EMPLOYEES(i).COMMISSION_PCT, V_EMPLOYEES(i).MANAGER_ID, V_EMPLOYEES(i).DEPARTMENT_ID, V_EMPLOYEES(i).BIRTHDAY;

if mod(i, v_rows) = 0 then

commit;

end if;

end loop;

commit;

end;

2.使用BULK COLLECT，不使用FORALL, 分批插入,多次提交,比较适用于大表;

create or replace procedure cp_data5 as

type t_cur is REF cursor;

c_table t_cur;

type t_employee is table of employees%rowtype;

v_employees t_employee;

rows number := 50;

v_sql varchar2(300);

v_table varchar(50);

begin

v_table := 'employee_cp';

open c_table for

select * from employees; --sour

v_sql := 'insert /*+ APPEND*/ into ' || v_table ||

' (EMPLOYEE_ID,

FIRST_NAME,

LAST_NAME,

EMAIL,

PHONE_NUMBER,

HIRE_DATE,

JOB_ID,

SALARY,

COMMISSION_PCT,

MANAGER_ID,

DEPARTMENT_ID,

BIRTHDAY) values (:1, :2,:3,:4,:5,:6, :7, :8,:9,:10, :11,:12)';

loop

fetch c_table bulk collect

into v_employees limit rows; --分批

dbms_output.put_line(v_employees.count);

for i in 1 .. v_employees.count loop

execute immediate v_sql

using V_EMPLOYEES(i).EMPLOYEE_ID, V_EMPLOYEES(i).FIRST_NAME, V_EMPLOYEES(i).LAST_NAME, V_EMPLOYEES(i).EMAIL, V_EMPLOYEES(i).PHONE_NUMBER, V_EMPLOYEES(i).HIRE_DATE, V_EMPLOYEES(i).JOB_ID, V_EMPLOYEES(i).SALARY, V_EMPLOYEES(i).COMMISSION_PCT, V_EMPLOYEES(i).MANAGER_ID, V_EMPLOYEES(i).DEPARTMENT_ID, V_EMPLOYEES(i).BIRTHDAY;

end loop;

commit;

exit when c_table%notfound;

end loop;

close c_table;

end;

3.使用BULK COLLECT和FORALL ，分批插入,多次提交,比较适用于大表; 前期数据字段定义比较烦锁(表各个字段必须分开定义)

-------------------

create or replace procedure cp_data as

type type_EMPLOYEE_ID is table of EMPLOYEES.EMPLOYEE_ID%type;

type type_FIRST_NAME is table of EMPLOYEES.FIRST_NAME%type;

type type_LAST_NAME is table of EMPLOYEES.LAST_NAME%type;

type type_EMAIL is table of EMPLOYEES.EMAIL%type;

type type_PHONE_NUMBER is table of EMPLOYEES.PHONE_NUMBER%type;

type type_HIRE_DATE is table of EMPLOYEES.HIRE_DATE%type;

type type_JOB_ID is table of EMPLOYEES.JOB_ID%type;

type type_SALARY is table of EMPLOYEES.SALARY%type;

type type_COMMISSION_PCT is table of EMPLOYEES.COMMISSION_PCT%type;

type type_MANAGER_ID is table of EMPLOYEES.MANAGER_ID%type;

type type_DEPARTMENT_ID is table of EMPLOYEES.DEPARTMENT_ID%type;

type type_BIRTHDAY is table of EMPLOYEES.BIRTHDAY%type;

V_EMPLOYEE_ID TYPE_EMPLOYEE_ID;

V_FIRST_NAME TYPE_FIRST_NAME;

V_LAST_NAME TYPE_LAST_NAME;

V_EMAIL TYPE_EMAIL;

V_PHONE_NUMBER TYPE_PHONE_NUMBER;

V_HIRE_DATE TYPE_HIRE_DATE;

V_JOB_ID TYPE_JOB_ID;

V_SALARY TYPE_SALARY;

V_COMMISSION_PCT TYPE_COMMISSION_PCT;

V_MANAGER_ID TYPE_MANAGER_ID;

V_DEPARTMENT_ID TYPE_DEPARTMENT_ID;

V_BIRTHDAY TYPE_BIRTHDAY;

type t_cur is ref cursor;

c_table t_cur;

v_table varchar2(30); --dest table

v_sql varchar2(300);

v_rows number := 50;

begin

v_table := 'EMPLOYEE_CP';

open c_table for

select * from employees; --sour table

v_sql := 'insert /*+ APPEND*/ into ' || v_table ||

' (EMPLOYEE_ID,

FIRST_NAME,

LAST_NAME,

EMAIL,

PHONE_NUMBER,

HIRE_DATE,

JOB_ID,

SALARY,

COMMISSION_PCT,

MANAGER_ID,

DEPARTMENT_ID,

BIRTHDAY)

values (:1, :2,:3,:4,:5,:6, :7, :8,:9,:10, :11,:12)';

loop

fetch c_table --.EMPLOYEE_ID, c_table.FIRST_NAME, c_table.LAST_NAME, c_table.EMAIL, c_table.PHONE_NUMBER, c_table.HIRE_DATE, c_table.JOB_ID, c_table.SALARY, c_table.COMMISSION_PCT, c_table.MANAGER_ID, c_table.DEPARTMENT_ID, c_table.BIRTHDAY

bulk collect

into V_EMPLOYEE_ID, V_FIRST_NAME, V_LAST_NAME, V_EMAIL, V_PHONE_NUMBER, V_HIRE_DATE, V_JOB_ID, V_SALARY, V_COMMISSION_PCT, V_MANAGER_ID, V_DEPARTMENT_ID, V_BIRTHDAY limit v_rows; --分批

forall i in 1 .. V_EMPLOYEE_ID.count execute immediate v_sql using

V_EMPLOYEE_ID(i), V_FIRST_NAME(i), V_LAST_NAME(i),

V_EMAIL(i), V_PHONE_NUMBER(i), V_HIRE_DATE(i),

V_JOB_ID(i), V_SALARY(i), V_COMMISSION_PCT(i),

V_MANAGER_ID(i), V_DEPARTMENT_ID(i), V_BIRTHDAY(i)

;

commit;

exit when c_table%notfound;

end loop;

end;

---------------------------------------------------------

4相关附助SQL:

select 'type TYPE_' || column_name || ' is table of ' || table_name || '.' ||

column_name || '%type'

from dba_tab_columns

where table_name = 'EMPLOYEES'

and owner = 'HYF'

select 'V_' || column_name || ' TYPE_' || column_name ||';'

from dba_tab_columns

where table_name = 'EMPLOYEES'

and owner = 'HYF'

select 'V_' || column_name || ','

from dba_tab_columns

where table_name = 'EMPLOYEES'

and owner = 'HYF'

select 'V_' || column_name || '(i),'

from dba_tab_columns

where table_name = 'EMPLOYEES'

and owner = 'HYF'