批量插入数据时主键冲突的处理

最新推荐文章于 2023-05-19 14:39:45 发布

congse3359

最新推荐文章于 2023-05-19 14:39:45 发布

阅读量1.2k

点赞数

客户有这样一个需求：有一个大表A，有主键，经常需要往这个表中批量插入大量数据，但插入的数据可能自身重复或者跟表A重复。

下面测试几种插入的方法：

DB:ORACLE 11.2.0.4

新建测试表：

新建表scott.tb_01

create table scott.tb_01

SELECT level c1,level c2,level c3

FROM DUAL

CONNECT BY LEVEL <= 100000;

在表scott.tb_01上增加主键

create unique index scott.pk_tb_01 on scott.tb_01(c1); --这一步不是必需的，因为下一步会自动建索引，但这种建索引再加主键的方式可以在建索引的时候加并行。

alter table scott.tb_01 add constraint pk_tb_01 primary key (c1) using index;

新建表scott.tb_02

create table scott.tb_02

SELECT level+100000 c1,level c2,level c3

FROM DUAL

CONNECT BY LEVEL <= 100000;

在表scott.tb_02上插入表scott.tb_01的100条数据做为重复数据

insert into scott.tb_02

SELECT c1,c2, c3

FROM scott.tb_01

where rownum <= 100;

commit;

现在需要将表scott.tb_02中和表scott.tb_01非重复的数据插入到scott.tb_01

方法1：关联插入

INSERT INTO SCOTT.TB_01

SELECT A.* --这里如果表SCOTT.TB_02自身有重复数据，还要加上distinct

FROM SCOTT.TB_02 A

LEFT JOIN SCOTT.TB_01 B

ON A.c1 = B.c2

WHERE B.c1 IS NULL;

100000 rows created.

上面一种常见的插入方法，这个方法的的问题在于，如果表SCOTT.TB_02和表SCOTT.TB_01都很大，两个表关联查询的成本会很高，影响性能。

方法2：使用HINT:IGNORE_ROW_ON_DUPKEY_INDEX

不使用HINT:

INSERT

INTO SCOTT.TB_01

SELECT * FROM SCOTT.TB_02 ;

ERROR at line 1:

ORA-00001: unique constraint (SCOTT.PK_TB_01) violated

从上面可以看出直接将表SCOTT.TB_02插入到表SCOTT.TB_01会出现主键冲突，报ORA-00001错误。

使用HINT:

INSERT /*+IGNORE_ROW_ON_DUPKEY_INDEX(a PK_tb_01)*/

INTO SCOTT.TB_01 A

SELECT * FROM SCOTT.TB_02 ;

100000 rows created.

从上面可以看出，加上HINT后，表SCOTT.TB_02中和表SCOTT.TB_01没有主键冲突的记录插入到表SCOTT.TB_01中，冲突的100条记录没有插入，也没有报错。

方法3：使用impdp的skip_constraint_errors选项

创建dump目录

create directory dump_home as '/home/oracle';

导出表scott.tb_02

expdp system tables=scott.tb_02 directory=dump_home dumpfile=expdp_tb_02.dmp logfile=expdp_tb_02.log

使用expdp导出表scott.tb_02，用于后续导入到表scott.tb_01中。

导入表scott.tb_02到scott.tb_01，不加skip_constraint_errors选项

impdp system tables=scott.tb_02 remap_table=scott.tb_02:tb_01 content=data_only directory=dump_home dumpfile=expdp_tb_02.dmp logfile=impdp_tb_02.log

Import: Release 11.2.0.4.0 - Production on Tue Dec 5 23:18:49 2017

Password:

Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production

With the Partitioning, OLAP, Data Mining and Real Application Testing options

Master table "SYSTEM"."SYS_IMPORT_TABLE_01" successfully loaded/unloaded

Starting "SYSTEM"."SYS_IMPORT_TABLE_01": system/******** tables=scott.tb_02 remap_table=scott.tb_02:tb_01 content=data_only directory=dump_home dumpfile=expdp_tb_02.dmp logfile=impdp_tb_02.log

Processing object type TABLE_EXPORT/TABLE/TABLE_DATA

ORA-31693: Table data object "SCOTT"."TB_01" failed to load/unload and is being skipped due to error:

ORA-00001: unique constraint (SCOTT.PK_TB_01) violated

Job "SYSTEM"."SYS_IMPORT_TABLE_01" completed with 1 error(s) at Tue Dec 5 23:19:02 2017 elapsed 0 00:00:09

从上面可以看出，如果不加skip_constraint_errors选项就会报主键冲突错误，导入失败。

导入表scott.tb_02到scott.tb_01，加skip_constraint_errors选项

impdp system tables=scott.tb_02 remap_table=scott.tb_02:tb_01 content=data_only directory=dump_home dumpfile=expdp_tb_02.dmp logfile=impdp_tb_02.log data_options=skip_constraint_errors

Import: Release 11.2.0.4.0 - Production on Tue Dec 5 23:21:29 2017

Password:

Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production

With the Partitioning, OLAP, Data Mining and Real Application Testing options

Master table "SYSTEM"."SYS_IMPORT_TABLE_01" successfully loaded/unloaded

Starting "SYSTEM"."SYS_IMPORT_TABLE_01": system/******** tables=scott.tb_02 remap_table=scott.tb_02:tb_01 content=data_only directory=dump_home dumpfile=expdp_tb_02.dmp logfile=impdp_tb_02.log data_options=skip_constraint_errors

Processing object type TABLE_EXPORT/TABLE/TABLE_DATA

. . imported "SCOTT"."TB_01" 1.796 MB 100000 out of 100100 rows

100 row(s) were rejected with the following error:

ORA-00001: unique constraint (SCOTT.PK_TB_01) violated

Rejected rows with the primary keys are:

Rejected row #1:

column C1: 1

…

Rejected row #99:

column C1: 99

Rejected row #100:

column C1: 100