Oracle使用fy_recover_data恢复truncate删除的数据

(一)truncate操作概述

在生产中，truncate是使用较多的命令，在使用不当的情况下，往往会造成表的数据全部丢失，恢复较为困难。对于truncate恢复，常见的有以下几种方法可以进行恢复：

使用数据泵导入。该方法操作简单，前提是必须要有备份可用，并且会有数据的丢失；
使用RMAN进行不完全恢复。可将数据库恢复到truncate之前的时刻，但是恢复时间较长；
使用odu、prm-dul、GDUL等收费软件进行恢复；
使用fy_recover_data包；

使用RMAN进行异机恢复已在之前测试过，详见：https://www.cnblogs.com/lijiaman/p/11577001.html 。

（二）FY_Recover_Data介绍

FY_Recover_Data是国内Oracle ACE大佬黄玮（个人网站：http://www.hellodba.com）开发的一个package，该脚本专门用于对truncate的表进行恢复。

根据作者所述，其原理：如果我们已经有一套元数据及数据块，然后将被TRUNCATE的用户数据块的内容取代其用户数据块的内容，是否可以“骗”过Oracle，让它读出这些数据呢？
回顾一下表扫描的过程，这个方法应该是可行的。我们只要想办法构造出一个结构相同、且具有完整元数据信息和格式化了的用户数据块的傀儡表对象，然后将被TRUNCATE的用户数据块找出，再将其数据内容部分嫁接到傀儡对象的用户数据块，使Oracle以外这是傀儡对象的数据，就能让Oracle扫描并读出数据内容。其原理用图示描述如下：

                                                +-------------------------+  
                                                | Copy Of Dummy Data File |  
                                                |  (With Formmated Blocks)|  
                                                +-------------------------+  
                                                            ||  
                                                            \/  
                                                (Blcok Header, Block Tail)  
                                                            ||  
                                                            \/  
+-------------------+                                +----------------+     Table Scan    +---------------+  
| Source Data File  | => (Data Block Content) =>     |  Dummy Table   |    ============>  | Restore Table |  
|(Without Meta Data)|                                |(With Meta Data)|                   +---------------+  
+-------------------+                                +----------------+

FY_Recover_Data对于表恢复的支持性如下：

压缩表	支持
索引组织表	支持
分区表	支持
行链接/行迁移	不支持
标准SQL类型	支持
BLOB/CLOB	支持Store in Row
离线恢复	支持
操作系统平台	全部
数据库版本	9i以上

本文使用FY_Recover_Data对truncate的几种情况进行恢复测试，以验证fy_recover_data的恢复能力。

(三)操作过程记录

（3.1）使用fy_recover_data包执行truncate恢复，truncate后未有新数据进入表

STEP1：创建测试表，并执行truncate

SQL> create table test01 as select * from dba_objects;

SQL> select count(*) from test01;

  COUNT(*)
----------
     86968

SQL> 
SQL> truncate table test01;

Table truncated

SQL> select count(*) from test01;

  COUNT(*)
----------
         0

STEP2：导入FY_Recover_Data.pck包

[oracle@source-node ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.4.0 Production on Tue Apr 21 10:50:17 2020

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

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

SQL> @/home/oracle/FY_Recover_Data.pck          -- 第一次执行发现第30行存在“&”符号，删除该符号

Enter value for files: 
old  30:   -- 1. Temp Restore and Recover tablespace & files      ---
new  30:   -- 1. Temp Restore and Recover tablespace       ---

Package created.

Warning: Package Body created with compilation errors.

SQL> @/home/oracle/FY_Recover_Data.pck      -- 删除“&”符号后导入成功

Package created.

Package body created.

STEP3：开始执行恢复，只需要两个参数：schema和table_name

[oracle@source-node ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.4.0 Production on Tue Apr 21 11:11:20 2020

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

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

SQL> set time on 
11:11:43 SQL> set serveroutput on
11:11:54 SQL> exec fy_recover_data.recover_truncated_table('LIJIAMAN','TEST01');
11:12:01: Use existing Directory Name: FY_DATA_DIR
11:12:02: Recover Table: LIJIAMAN.TEST01$
11:12:02: Restore Table: LIJIAMAN.TEST01$$
11:12:09: Copy file of Recover Tablespace: FY_REC_DATA_COPY.DAT1
11:12:09: begin to recover table LIJIAMAN.TEST01
11:12:09: Use existing Directory Name: TMP_HF_DIR
11:12:09: Recovering data in datafile
/u01/app/oracle/oradata/testdb1/users01.dbf
11:12:09: Use existing Directory Name: TMP_HF_DIR
11:12:39: 1242 truncated data blocks found.
11:12:39: 86968 records recovered in backup table LIJIAMAN.TEST01$$
11:12:39: Total: 1242 truncated data blocks found.
11:12:39: Total: 86968 records recovered in backup table LIJIAMAN.TEST01$$
11:12:39: Recovery completed.
11:12:39: Data has been recovered to LIJIAMAN.TEST01$$

PL/SQL procedure successfully completed.

STEP4：根据恢复日志，会创建临时中转表test01$和test01$$，恢复的数据保存在test01$$中

SQL> show user
User is "LIJIAMAN"

SQL> select count(*) from test01$$;

  COUNT(*)
----------
     86968

--将数据还原到test01表中

SQL> insert into test01 select * from test01$$;

--确认数据已经还原回来

SQL> select count(*) from test01;

  COUNT(*)
----------
     86968

经过测试，如果表被truncate后，未执行其它操作，数据可以使用fy_recover_data恢复回来。

#######################################################################################

（3.2）使用fy_recover_data包执行truncate恢复，truncate后有新数据进入表（新插入的数据比truncate之前多）

STEP1：创建测试表、序列、存储过程

SQL> create table test01
  2  (
  3      col1      number,
  4      col2      number,
  5      col3      date,
  6      col4      varchar2(30),
  7      col5      varchar2(100)
  8  );

Table created
SQL> --创建自增序列
SQL> CREATE SEQUENCE seq01
  2    START WITH 1
  3    MAXVALUE 99999999
  4    MINVALUE 0
  5    CYCLE
  6    CACHE 10
  7    ORDER;

Sequence created

 
SQL> --创建随机数据插入存储过程,其中col1列单调递增
create or replace procedure p_insert_test01 IS
v_col1 NUMBER;
BEGIN
FOR i IN 1..10000 LOOP
select seq01.nextval INTO v_col1 from dual;
insert into test01(col1,col2,col3,col4,col5)
values
(v_col1,
(select round(dbms_random.value(10000, 100000000)) from dual),
sysdate,
(select dbms_random.string('a', 25) from dual),
(select dbms_random.string('a', 85) from dual));
END LOOP;
commit;
end p_insert_test01;

STEP2：测试表插入10000条数据，col1列的值从1到10000

SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

SQL> select count(*) from test01;

  COUNT(*)
----------
     10000

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
         1      10000

STEP3：执行truncate操作

SQL> truncate table test01;

Table truncated

STEP4：接着往表里插入20000条数据

SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

SQL> select count(*) from test01;

  COUNT(*)
----------
     20000

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
     10001      30000

STEP5：执行恢复操作

[oracle@source-node ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.4.0 Production on Tue Apr 21 14:00:57 2020

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

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

SQL> set serveroutput on
SQL> set time on
14:01:06 SQL> 
14:01:09 SQL> exec fy_recover_data.recover_truncated_table('LIJIAMAN','TEST01');
14:01:13: Use existing Directory Name: FY_DATA_DIR
14:01:13: Recover Table: LIJIAMAN.TEST01$
14:01:14: Restore Table: LIJIAMAN.TEST01$$
14:01:18: Copy file of Recover Tablespace: FY_REC_DATA_COPY.DAT1
14:01:18: begin to recover table LIJIAMAN.TEST01
14:01:18: Use existing Directory Name: TMP_HF_DIR
14:01:18: Recovering data in datafile
/u01/app/oracle/oradata/testdb1/users01.dbf
14:01:18: Use existing Directory Name: TMP_HF_DIR
14:01:32: 402 truncated data blocks found.
14:01:32: 20000 records recovered in backup table LIJIAMAN.TEST01$$
14:01:32: Total: 402 truncated data blocks found.
14:01:32: Total: 20000 records recovered in backup table LIJIAMAN.TEST01$$
14:01:32: Recovery completed.
14:01:32: Data has been recovered to LIJIAMAN.TEST01$$

PL/SQL procedure successfully completed.

STEP6：通过对test01$$表进行确认，发现返回的数据是truncate之后插入的数据，不符合要求

SQL> select count(*) from test01;

  COUNT(*)
----------
     20000
SQL> select count(*) from test01$$;

  COUNT(*)
----------
     20000

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
     10001      30000

SQL> SELECT MIN(col1),MAX(col1) FROM test01$$;

MIN(COL1)  MAX(COL1)
---------- ----------
     10001      30000

#######################################################################################

（3.3）使用fy_recover_data包执行truncate恢复，truncate后有新数据进入表（新插入的数据比truncate之前少）

STEP1：创建测试表、序列、存储过程

Connected to Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 
Connected as lijiaman@192.168.10.11/testdb1
SQL> DROP TABLE  test01 PURGE;

Table dropped
SQL> create table test01
  2  (
  3  col1 number,
  4  col2 number,
  5  col3 date,
  6  col4 varchar2(30),
  7  col5 varchar2(100)
  8  );

Table created
SQL> DROP SEQUENCE seq01;

Sequence dropped
SQL> --创建自增序列
SQL> CREATE SEQUENCE seq01
  2  START WITH 1
  3  MAXVALUE 99999999
  4  MINVALUE 0
  5  CYCLE
  6  CACHE 10
  7  ORDER;

Sequence created
SQL> --创建随机数据插入存储过程,其中col1列单调递增
SQL> create or replace procedure p_insert_test01 IS
  2  v_col1 NUMBER;
  3  BEGIN
  4  FOR i IN 1..10000 LOOP
  5  select seq01.nextval INTO v_col1 from dual;
  6  insert into test01(col1,col2,col3,col4,col5)
  7  values
  8  (v_col1,
  9  (select round(dbms_random.value(10000, 100000000)) from dual),
 10  sysdate,
 11  (select dbms_random.string('a', 25) from dual),
 12  (select dbms_random.string('a', 85) from dual));
 13  END LOOP;
 14  commit;
 15  end p_insert_test01;

16  /

Procedure created

STEP2：测试表插入10000条数据，col1列的值从1到10000

SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

SQL> select count(*) from test01;

  COUNT(*)
----------
     10000

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
         1      10000

STEP3：执行truncate操作

SQL> truncate table test01;

Table truncated

STEP4：修改存储过程，酶促插入100条数据

SQL> --创建随机数据插入存储过程,其中col1列单调递增
SQL> create or replace procedure p_insert_test01 IS
  2  v_col1 NUMBER;
  3  BEGIN
  4  FOR i IN 1..100 LOOP
  5  select seq01.nextval INTO v_col1 from dual;
  6  insert into test01(col1,col2,col3,col4,col5)
  7  values
  8  (v_col1,
  9  (select round(dbms_random.value(10000, 100000000)) from dual),
10  sysdate,
11  (select dbms_random.string('a', 25) from dual),
12  (select dbms_random.string('a', 85) from dual));
 13  END LOOP;
 14  commit;
 15  end p_insert_test01;
 16  /

Procedure created

-- 测试表插入100条数据
SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

STEP5：执行恢复操作

[oracle@source-node ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.4.0 Production on Tue Apr 21 14:22:34 2020

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

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

SQL> set time on
14:22:39 SQL> set serveroutput on
14:22:44 SQL> exec fy_recover_data.recover_truncated_table('LIJIAMAN','TEST01');
14:22:52: Use existing Directory Name: FY_DATA_DIR
14:22:52: Recover Table: LIJIAMAN.TEST01$
14:22:52: Restore Table: LIJIAMAN.TEST01$$
14:22:57: Copy file of Recover Tablespace: FY_REC_DATA_COPY.DAT1
14:22:57: begin to recover table LIJIAMAN.TEST01
14:22:57: Use existing Directory Name: TMP_HF_DIR
14:22:57: Recovering data in datafile
/u01/app/oracle/oradata/testdb1/users01.dbf
14:22:57: Use existing Directory Name: TMP_HF_DIR
14:23:06: 5 truncated data blocks found.
14:23:06: 100 records recovered in backup table LIJIAMAN.TEST01$$
14:23:06: Total: 5 truncated data blocks found.
14:23:06: Total: 100 records recovered in backup table LIJIAMAN.TEST01$$
14:23:06: Recovery completed.
14:23:06: Data has been recovered to LIJIAMAN.TEST01$$

PL/SQL procedure successfully completed.

STEP6：通过对test01$$表进行确认，发现返回的数据是truncate之后插入的数据，不符合要求

SQL> select count(*) from test01;

  COUNT(*)
----------
       100

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
     10001      10100

SQL> 
SQL> select count(*) from test01$$;

  COUNT(*)
----------
       100

SQL> SELECT MIN(col1),MAX(col1) FROM test01$$;

MIN(COL1)  MAX(COL1)
---------- ----------
     10001      10100

#######################################################################################
（3.4）测试数据文件被覆盖是否影响恢复

STEP1：创建测试表

Connected to Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 
Connected as lijiaman@192.168.10.11/testdb1

SQL> create table test01
  2  (
  3  col1 number,
  4  col2 number,
  5  col3 date,
  6  col4 varchar2(30),
  7  col5 varchar2(100)
  8  ) TABLESPACE USERS;

Table created

STEP2：初始时候，表空间总共20MB，剩余15.94MB

SQL> SELECT  SUBSTR(a.TABLESPACE_NAME,1,30) TablespaceName,
  2      round(SUM(a.bytes/1024/1024),2)  AS "Totle_size(MB)",
  3      round(SUM(NVL(b.free_space1/1024/1024,0)),2) AS "Free_space(MB)",
  4      round(SUM(a.bytes/1024/1024),2)-round(SUM(NVL(b.free_space1/1024/1024,0)),2)  AS "Used_space(MB)",
  5      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0))) *100/SUM(a.bytes/1024/1024),2) AS "Used_percent%",
  6      round(SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2)                AS "Max_size(MB)",
  7      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0)))*100/SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2) AS "Max_percent%"
  8    FROM dba_data_files a,
  9      (SELECT SUM(NVL(bytes,0)) free_space1,
 10        file_id
 11      FROM dba_free_space
 12    GROUP BY file_id
 13     ) b
 14  WHERE a.file_id = b.file_id(+)
 15  AND   a.TABLESPACE_NAME = 'USERS'
 16   GROUP BY a.TABLESPACE_NAME;

TABLESPACENAME                                                                   Totle_size(MB) Free_space(MB) Used_space(MB) Used_percent% Max_size(MB) Max_percent%
-------------------------------------------------------------------------------- -------------- -------------- -------------- ------------- ------------ ------------
USERS                                                                                        20          15.94           4.06         20.31           20        20.31

STEP3：test01表插入大量数据

SQL> exec p_insert_test01;
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
PL/SQL procedure successfully completed

SQL> /
begin p_insert_test01; end;

ORA-01653: unable to extend table LIJIAMAN.TEST01 by 128 in tablespace USERS
ORA-06512: at "LIJIAMAN.P_INSERT_TEST01", line 6
ORA-06512: at line 1

STEP4：此时，表空间总共20MB，剩余0.94MB

SQL> SELECT  SUBSTR(a.TABLESPACE_NAME,1,30) TablespaceName,
  2      round(SUM(a.bytes/1024/1024),2)  AS "Totle_size(MB)",
  3      round(SUM(NVL(b.free_space1/1024/1024,0)),2) AS "Free_space(MB)",
  4      round(SUM(a.bytes/1024/1024),2)-round(SUM(NVL(b.free_space1/1024/1024,0)),2)  AS "Used_space(MB)",
  5      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0))) *100/SUM(a.bytes/1024/1024),2) AS "Used_percent%",
  6      round(SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2)                AS "Max_size(MB)",
  7      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0)))*100/SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2) AS "Max_percent%"
  8    FROM dba_data_files a,
  9      (SELECT SUM(NVL(bytes,0)) free_space1,
 10        file_id
 11      FROM dba_free_space
 12    GROUP BY file_id
 13     ) b
 14  WHERE a.file_id = b.file_id(+)
 15  AND   a.TABLESPACE_NAME = 'USERS'
 16   GROUP BY a.TABLESPACE_NAME;

TABLESPACENAME                                                                   Totle_size(MB) Free_space(MB) Used_space(MB) Used_percent% Max_size(MB) Max_percent%
-------------------------------------------------------------------------------- -------------- -------------- -------------- ------------- ------------ ------------
USERS                                                                                        20           0.94          19.06         95.31           20        95.31

STEP5：此时test01表有90000行数据

SQL> select count(*) from test01;

  COUNT(*)
----------
     90000

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------
    109751     199750

STEP6：对test01执行truncate

SQL> truncate table test01;

Table truncated

STEP7：执行truncate后，空间已经释放

SQL> SELECT  SUBSTR(a.TABLESPACE_NAME,1,30) TablespaceName,
  2      round(SUM(a.bytes/1024/1024),2)  AS "Totle_size(MB)",
  3      round(SUM(NVL(b.free_space1/1024/1024,0)),2) AS "Free_space(MB)",
  4      round(SUM(a.bytes/1024/1024),2)-round(SUM(NVL(b.free_space1/1024/1024,0)),2)  AS "Used_space(MB)",
  5      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0))) *100/SUM(a.bytes/1024/1024),2) AS "Used_percent%",
  6      round(SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2)                AS "Max_size(MB)",
  7      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0)))*100/SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2) AS "Max_percent%"
  8    FROM dba_data_files a,
  9      (SELECT SUM(NVL(bytes,0)) free_space1,
 10        file_id
 11      FROM dba_free_space
 12    GROUP BY file_id
 13     ) b
 14  WHERE a.file_id = b.file_id(+)
 15  AND   a.TABLESPACE_NAME = 'USERS'
 16   GROUP BY a.TABLESPACE_NAME;

TABLESPACENAME                                                                   Totle_size(MB) Free_space(MB) Used_space(MB) Used_percent% Max_size(MB) Max_percent%
-------------------------------------------------------------------------------- -------------- -------------- -------------- ------------- ------------ ------------
USERS                                                                                        20          15.88           4.12         20.63           20        20.63

STEP8：创建表test02，用来覆盖test01释放的空间

SQL> create table test02 as select * from dba_objects;

Table created

STEP9：test02表创建之后，剩余空间为5.88MB，可以说明：test02表的数据占用了test01表释放出来的空间，即test01表的部分数据已经被覆盖

SQL> SELECT  SUBSTR(a.TABLESPACE_NAME,1,30) TablespaceName,
  2      round(SUM(a.bytes/1024/1024),2)  AS "Totle_size(MB)",
  3      round(SUM(NVL(b.free_space1/1024/1024,0)),2) AS "Free_space(MB)",
  4      round(SUM(a.bytes/1024/1024),2)-round(SUM(NVL(b.free_space1/1024/1024,0)),2)  AS "Used_space(MB)",
  5      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0))) *100/SUM(a.bytes/1024/1024),2) AS "Used_percent%",
  6      round(SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2)                AS "Max_size(MB)",
  7      ROUND((SUM(a.bytes/1024/1024)-SUM(NVL(b.free_space1/1024/1024,0)))*100/SUM((case when a.MAXBYTES = 0 then a.bytes else a.MAXBYTES end)/1024/1024),2) AS "Max_percent%"
  8    FROM dba_data_files a,
  9      (SELECT SUM(NVL(bytes,0)) free_space1,
 10        file_id
 11      FROM dba_free_space
 12    GROUP BY file_id
 13     ) b
 14  WHERE a.file_id = b.file_id(+)
 15  AND   a.TABLESPACE_NAME = 'USERS'
 16   GROUP BY a.TABLESPACE_NAME;

TABLESPACENAME                                                                   Totle_size(MB) Free_space(MB) Used_space(MB) Used_percent% Max_size(MB) Max_percent%
-------------------------------------------------------------------------------- -------------- -------------- -------------- ------------- ------------ ------------
USERS                                                                                        20           5.88          14.12         70.63           20        70.63

STEP10：执行恢复操作

[oracle@source-node ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.4.0 Production on Tue Apr 21 15:09:58 2020

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

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

SQL> set time on
15:10:05 SQL> set serveroutput on
15:10:10 SQL> exec fy_recover_data.recover_truncated_table('LIJIAMAN','TEST01');
15:10:17: Use existing Directory Name: FY_DATA_DIR
15:10:17: Recover Table: LIJIAMAN.TEST01$
15:10:17: Restore Table: LIJIAMAN.TEST01$$
15:10:22: Copy file of Recover Tablespace: FY_REC_DATA_COPY.DAT1
15:10:22: begin to recover table LIJIAMAN.TEST01
15:10:22: Use existing Directory Name: TMP_HF_DIR
15:10:22: Recovering data in datafile
/u01/app/oracle/oradata/testdb1/users01.dbf
15:10:22: Use existing Directory Name: TMP_HF_DIR
15:10:31: 645 truncated data blocks found.
15:10:31: 24439 records recovered in backup table LIJIAMAN.TEST01$$
15:10:31: Total: 645 truncated data blocks found.
15:10:31: Total: 24439 records recovered in backup table LIJIAMAN.TEST01$$
15:10:31: Recovery completed.
15:10:31: Data has been recovered to LIJIAMAN.TEST01$$

PL/SQL procedure successfully completed.

STEP11：发现只恢复了部分数据，不符合要求

-- truncate之前test01表有90000行数据，恢复了24339行数据

SQL> select count(*) from test01$$;

  COUNT(*)
----------
     24439

SQL> SELECT MIN(col1),MAX(col1) FROM test01;

MIN(COL1)  MAX(COL1)
---------- ----------

SQL> 
SQL> SELECT MIN(col1),MAX(col1) FROM test01$$;

MIN(COL1)  MAX(COL1)
---------- ----------
    109751     199750

(四)总结

对于使用工具fy_recover_data进行数据恢复，需要确保：

①truncate之后，需要保证没有新的数据进入表中，否则无法还原；

②存放该表的数据文件块不能被覆盖，否则无法完整还原数据。

在发生故障后，可以迅速使用：

SQL> alter tablespace users read only;
SQL> alter tablespace users read write;

来关闭/开启表空间的写功能，这样可以保证数据文件不会被覆写。

【完】

来源：oschina

链接：https://my.oschina.net/u/4389867/blog/3301174

标签

records

Sequence

sqlplus

Dat