1. Backup and Recovery User's Guide
  2. Diagnosing and Responding to Failures
  3. Recovering Tables and Table Partitions

22 Recovering Tables and Table Partitions

This chapter describes how to recover tables and table partitions to a specified point in time. This chapter contains the following topics:

22.1 Overview of Recovering Tables and Table Partitions

The RMAN RECOVER command enables you to recover tables and table partitions from RMAN backups.

Note:

There are other methods of recovering tables to a specified point in time such as Oracle Flashback and TSPITR. For more information about the scenarios in which these methods are useful and how to recover tables using these methods, see:

See Also:

22.1.1 Purpose of Recovering Tables and Table Partitions from RMAN Backups

RMAN enables you to recover one or more tables or table partitions to a specified point in time without affecting the remaining database objects. You can use previously-created RMAN backups to recover tables and table partitions to a specified point in time.

Recovering tables and table partitions from RMAN backups is useful in the following scenarios:

  • You need to recover a very small number of tables to a particular point in time. In this situation, TSPITR is not the most effective solution because it moves all the objects in the tablespace to a specified point in time.

  • You need to recover tables that have been logically corrupted or have been dropped and purged.

  • Flashback Table is not possible because the desired point-in-time is older than available undo.

  • You want to recover data that is lost after a DDL operation modified the structure of tables. Using Flashback Table is not possible because a DDL was run on the tables between the desired point in time and the current time. Flashback Table cannot rewind tables through structural changes such as a truncate table operation.

22.1.2 RMAN Backups Required to Recover Tables and Table Partitions

To recover tables or table partitions, you need a full backup of undo, SYSTEM, SYSAUX, and the tablespace that contains the tables or table partitions.

To recover a table, all partitions that contain the dependent objects of the table must be included in the recovery set. If the indexes or partitions for a table in tablespace tbs1 are contained in tablespace tbs2, then you can recover the table only if tablepsace tbs2 is also included in the recovery set.

To recover tables in a PDB, you need backups of the following:

  • SYSTEM, SYSAUX, and undo tablespace of the root, CDB seed, and the PDB containing the tables

  • Tablespace containing the tables or partitions

See Also:

Prerequisites for Recovering Tables and Table Partitions from RMAN Backups

22.1.3 Basic Concepts of Recovering Tables and Table Partitions from RMAN Backups

RMAN uses the RECOVER command to recover tables or table partitions to a specified point in time.

To recover tables and table partitions from an RMAN backup, you need to provide the following information:

  • Names of tables or table partitions that must be recovered

  • Point in time to which the tables or table partitions must be recovered

  • Whether the recovered tables or table partitions must be imported into the target database

RMAN uses this information to automate the process of recovering the specified tables or table partitions. As part of the recovery process, RMAN creates an auxiliary database that is used to recover tables or table partitions to the specified point in time.

If the recovered tables or table partitions need to be renamed, mapped to a new tablespace, or mapped to a new schema, then you must specify the new names for the tables, tablespaces, or schemas.

See Also:

22.1.3.1 Steps Performed By RMAN to Recover Tables and Table Partitions

RMAN performs a series of steps while automating the process of recovering tables or table partitions from an RMAN backup.

The steps include the following:

  1. Determines which backup contains the tables or table partitions that need to be recovered, based on the point in time specified for the recovery.
  2. Determines if there is sufficient space on the target host to create the auxiliary instance that will be used during the table or partition recovery process.
    If the required space is not available, then RMAN displays an error and exits the recovery operation.
  3. Creates an auxiliary database on the target host and recovers the specified tables or table partitions, until the specified point in time, into this auxiliary database.

    You can specify the location on the target host to which the recovered data files are stored in the auxiliary database.

  4. Creates a Data Pump export dump file that contains the recovered tables or table partitions.

    You can specify the name and the location of the Data Pump export dump file used to store the metadata of the recovered tables or table partitions.

  5. (Optional) Imports the Data Pump export dump file into the target instance.

    You can choose not to import the export dump file that contains the recovered tables or table partitions into the target database. If you do not import the export dump file as part of the recovery process, you must manually import it later using the Data Pump Import utility.

  6. (Optional) Renames the recovered tables or table partitions in the target database.

    You can also import recovered objects into a tablespace or schema that is different from the one in which they originally existed.

See Also:

22.1.3.2 About the Location of Auxiliary Database Files During RMAN Table Recovery

RMAN creates an auxiliary database that it uses during the process of recovering the specified tables or table partitions. Multiple techniques are available to specify the location of auxiliary database files.

On the target host that is used to store data files for the auxiliary database, use one of the following techniques:

  • AUXILIARY DESTINATION clause in the RECOVER command

  • SET NEWNAME command

    Use a RUN block containing the RECOVER command and required SET NEWNAME commands that rename the data files.

See Also:

It is recommended that you provide a location for data files in the auxiliary database by using the AUXILIARY DESTINATION clause. When you use the SET NEWNAME command, if you omit the name of even one data file required for the recovery process, the tables or table partitions cannot be recovered.

22.1.3.3 About the Data Pump Export Dump File Used During RMAN Table Recovery

After recovering tables or table partitions to the specified point in time on the auxiliary database, RMAN creates a Data Pump export dump file that contains the recovered objects. You can either specify a name and location for this dump file or allow RMAN to use a default name and location.

Use the DATAPUMP DESTINATION clause of the RECOVER command to specify the location in which the Data Pump export dump file is created. The location is typically the path of the operating-system directory that stores the dump file. If you omit this clause, the dump file is stored in the location specified by the AUXILIARY DESTINATION parameter. If you do not specify an auxiliary destination, the dump file is stored in a default operating system-specific location. On Linux, this default location is $ORACLE_HOME/dbs. On Windows, the default location is %ORACLE_HOME\database.

Use the DUMP FILE clause of the RECOVER command to specify the name of the Data Pump export dump file. If you omit this clause, RMAN uses a default operating system-specific name for the dump file. On Linux and Windows, the default dump file name is tspitr_SID-of-clone_n.dmp, where SID-of-clone is the Oracle SID of the auxiliary database created by RMAN to perform the recovery and n is any randomly-generated number. If a file with the name specified by DUMP FILE exists in the location in which the dump file must be created, then the recover operation fails.

22.1.3.4 About Importing Recovered Tables and Table Partitions into the Target Database

By default, RMAN imports the recovered tables or table partitions, which are stored in the export dump file, into the target database. However, you can choose not to import the recovered tables or table partitions by using the NOTABLEIMPORT clause of the RESTORE command.

When NOTABLEIMPORT is used, RMAN recovers them to the specified point and then creates the export dump file. However, this dump file is not imported into the target database. You must manually import this dump file into your target database, when required, by using the Data Pump Import utility.

If an error occurs during the import operation, RMAN does not delete the export dump file at the end of the table recovery. This enables you to manually import the dump file.

22.1.3.5 About Renaming Recovered Tables and Table Partitions During RMAN Recovery

Use the REMAP TABLE clause to rename recovered tables or table partitions in the target database.

If a table with the same name as the one that you recovered exists in the target database, RMAN displays an error message indicating that the REMAP TABLE clause must be used to rename the recovered table.

When you recover table partitions, each table partition is recovered into a separate table. Use the REMAP TABLE clause to specify the table names into which each recovered partition must be imported. If you do not explicitly specify table names, RMAN generates table names by concatenating the recovered table name and partition name. The generated names are in the format tablename_partitionname. If a table with this name exists in the target database, then RMAN appends _1 to the name. If this name too exists, then RMAN appends _2 to the name and so on.

To import the recovered tables or table partitions into a tablespace that is different from the one in which these objects originally existed, use the REMAP TABLESPACE clause of the RECOVER command. Only the tables or table partitions that are being recovered are remapped, the existing objects are not changed.

Note:

When you use the REMAP option, any named constraints and indexes are not imported. This is to avoid name conflicts with existing tables.

22.1.3.6 About Recovering Tables and Partitions Into a New Schema

Recovering tables or table partitions into a different schema enables you to avoid name conflicts that may be caused by constraint, index, or trigger names that already existing in the source schema.

Starting with Oracle Database 12c Release 2 (12.2), you can recover tables or table partitions into a schema that is different from the source schema (the schema in which they originally existed). While recovering objects into a different schema, you can either retain their original names or rename them. You can rename tables and remap the schema in a single recovery operation. For example, you can recover the HR.EMPLOYEES table either into the NEW_HR.EMPLOYEES table, the HR.NEW_EMPLOYEES table, or the NEW_HR.NEW_EMPLOYEES table. The REMAP TABLE clause enables you to rename objects and recover them into a different schema.

During table recovery, use the REMAP TABLE clause of the RECOVER TABLE command to map the source schema to a new schema. The new schema must exist in the target database before you perform the recovery.

Note:

Table recovery is not supported on physical standby databases. For logical standby databases, objects that are recovered on the primary database are propagated to the logical standby.

Note:

Example: Recovering a Table into a New Schema

22.1.4 Limitations of Recovering Tables and Table Partitions from RMAN Backups

Recovering tables and table partitions from RMAN backups by using the RECOVER command is subject to certain limitations.

The limitations include the following:

  • Tables and table partitions belonging to SYS schema cannot be recovered.

  • Tables and table partitions from SYSTEM and SYSAUX tablespaces cannot be recovered.

  • Tables and table partitions on standby databases cannot be recovered.

  • Tables with named NOT NULL constraints cannot be recovered with the REMAP option.

22.2 Preparing to Recover Tables and Table Partitions

You must perform some preliminary tasks before you prepare to recover tables or table partitions .

Review the limitations described in "Limitations of Recovering Tables and Table Partitions from RMAN Backups".

The preparation for recovering tables or table partitions from RMAN backups includes the following steps:

  • Verifying that the prerequisites required to recover tables or table partitions are met

  • Determining the point in time to which the tables or table partitions must be recovered

  • Deciding if the recovered tables or table partitions must be imported into the target database

    By default, RMAN imports the recovered tables or table partitions into the target database. However, you can specify that RMAN must not import the recovered objects.

  • Deciding if the recovered tables or table partitions must be renamed, mapped to a new tablespace, or mapped to a new schema.

See Also:

22.2.1 Prerequisites for Recovering Tables and Table Partitions from RMAN Backups

Certain prerequisites must be met before you recover tables or table partitions from RMAN backups.

They include the following:

  • The target database must be in read-write mode.

  • The target database must be in ARCHIVELOG mode.

  • You must have RMAN backups of the tables or table partitions as they existed at the point in time to which you want recover these objects.

  • To recover single table partitions, the COMPATIBLE initialization parameter for target database must be set to 11.1.0 or higher.

See Also:

RMAN Backups Required to Recover Tables and Table Partitions

22.2.2 Determining the Point-in-time to Which Tables and Table Partitions Must be Recovered

It is important to determine the exact point in time to which you want to recover the tables or table partitions. There are multiple ways to specify the point in time to which objects must be recovered.

Use one of the following methods:

  • SCN

    Recovers tables or table partitions to the state that they were at the time specified by the SCN.

  • Time

    Recovers tables or table partitions to the state they were in at the specified time. Use the date format specified in the NLS_LANG and NLS_DATE_FORMAT environment variables. You can also use data constants such as SYSDATE to specify the time, for example SYSDATE-30.

  • Sequence number

    Recovers tables or table partitions to the state they were at the time specified by the log sequence number and thread number.

22.3 Recovering Tables and Table Partitions

Use the RESTORE and RECOVER commands to recover tables or table partitions.

To recover tables or table partitions in a non-CDB to a specified point in time:

  1. Perform the planning tasks described in "Preparing to Recover Tables and Table Partitions".
  2. Start RMAN and connect as TARGET to the target database. You must connect as a user with the SYSBACKUP or SYSDBA privilege, as described in "Making Database Connections with RMAN".
  3. Recover the selected tables or table partitions to the specified point in time by using the RECOVER TABLE command. You must use the AUXILIARY DESTINATION clause and one of the following clauses to specify the point in time for recovery: UNTIL TIME, UNTIL SCN, or UNTIL SEQUENCE.

    You can use the following additional clauses in the RECOVER command:

See Also:

Examples: Recovering Tables and Table Partitions From RMAN Backups

22.4 Recovering Tables and Table Partitions in PDBs

RMAN enables you to use the RECOVER command to recover one or more tables or table partitions in a pluggable database (PDB) to a specified point-in-time, without impacting other objects in the PDB.

The steps used to recover tables or table partitions in a PDB are similar to the ones used for non-CDBs, with the differences described in this section.

To recover tables or table partitions in a PDB:

  1. Perform the planning tasks described in "Preparing to Recover Tables and Table Partitions".
  2. Start RMAN and connect to the root as a user with the SYSDBA or SYSBACKUP privilege, as described in "Making RMAN Connections to a CDB".
  3. Recover the tables or table partitions to the specified point in time by using the RECOVER TABLE ... OF PLUGGABLE DATABASE command.

    You must use the AUXILIARY DESTINATION clause and one of the following clauses: UNITL TIME, UNTIL SCN, or UNTIL SEQUENCE.

    Depending on your requirements, you may also need to use the one or more of the following clauses: DUMP FILE, DATAPUMP DESTINATION, NOTABLEIMPORT, REMAP TABLE, or REMAP TABLESPACE.

    See Also:

    The following command recovers the table PDB_EMP in the PDB HR_PDB to the state that it was in 4 days before the current date. HR is the name of the schema that contains the table. The recovered table is renamed to EMP_RECVR. 

    RECOVER TABLE HR.PDB_EMP OF PLUGGABLE DATABASE HR_PDB
UNTIL TIME 'SYSDATE-4'
AUXILIARY DESTINATION '/tmp/backups'
REMAP TABLE 'HR'.'PDB_EMP':'EMP_RECVR';

22.5 Examples: Recovering Tables and Table Partitions From RMAN Backups

This section contains examples that cover multiple scenarios for recovering tables and table partitions.

22.5.1 Example: Recovering Tables to a Specified Point in Time

This example recovers multiple tables to a specified point in time that is represented using SYSDATE.

Assume that you want to recover two tables EMP and DEPT to the state they were in two days ago, before some logical corruption occurred. However, you do not want RMAN to import these tables into the target database. RMAN must only create the export dump file, called emp_dept_exp_dump.dat, in the location /tmp/recover/dumpfiles. Using NOTABLEIMPORT indicates that these tables must not be imported into the target database. You can import these tables, when required, by using the Data Pump import utility. The auxiliary destination used during the recovery process is /tmp/oracle/recover.

To recover tables EMP and DEPT without importing them into the target database:

  1. Perform the planning tasks described in "Preparing to Recover Tables and Table Partitions".

    In this example, you need to recover tables to a point in time specified by an expression that uses SYSDATE. However, the recovered tables must not be imported in to the target database.

  2. Start an RMAN session and connect as TARGET to the target database as described in "Making Database Connections with RMAN".
  3. Recover the tables EMP and DEPT using the following clauses in the RECOVER command: DATAPUMP DESTINATION, DUMP FILE, REMAP TABLE, and NOTABLEIMPORT.

    The following RECOVER command recovers the EMP and DEPT tables. 

    RECOVER TABLE SCOTT.EMP, SCOTT.DEPT
    UNTIL TIME 'SYSDATE-1'
    AUXILIARY DESTINATION '/tmp/oracle/recover'
    DATAPUMP DESTINATION '/tmp/recover/dumpfiles'
    DUMP FILE 'emp_dept_exp_dump.dat'
    NOTABLEIMPORT;

See Also:

Oracle Database Backup and Recovery Reference for additional examples about recovering tables to a specified point in time

22.5.2 Example: Recovering Table Partitions to a Specified Log Sequence Number

This example uses RMAN backups to recover multiple table partitions.

In this example, the table sales, in the schema sh, contains the following partitions: sales_1998, sales_1999, sales_2000, and sales_2001. This table is stored in the sales_ts tablespace. You need to recover two partitions, sales_1998 and sales_1999, to a point in time that is specified by a redo log sequence number. The recovered tables must be automatically imported into the target database and mapped to the tablespace SALES_PRE_2000_TS.

To recover the partitions sales_1998 and sales_1999 to a specified log sequence number:

  1. Perform the planning tasks described in "Preparing to Recover Tables and Table Partitions".

    In this example, you need to recover two table partitions to a specified log sequence number and then import these partitions into the target database.

  2. Start an RMAN session and connect as TARGET to the target database as described in "Making Database Connections with RMAN".
  3. Recover partitions using the following RECOVER command with the REMAP TABLE and REMAP TABLESPACE clauses.
    RECOVER TABLE SH.SALES:SALES_1998, SH.SALES:SALES_1999
    UNTIL SEQUENCE 354
    AUXILIARY DESTINATION '/tmp/oracle/recover'
    REMAP TABLE 'SH'.'SALES':'SALES_1998':'HISTORIC_SALES_1998',
              'SH'.'SALES':'SALES_1999':'HISTORIC_SALES_1999' 
    REMAP TABLESPACE 'SALES_TS':'SALES_PRE_2000_TS';

    In this case, the specified table partitions are imported as separate tables, called historic_sales_1998 and historic_sales_1999, into the sales_pre_2000_ts tablespace of the target database. The REMAP TABLE clause specifies the names used for the imported tables. The auxiliary destination used during the recovery process is /tmp/oracle/recover.

    If you omit the REMAP TABLE clause, RMAN uses default names for the imported tables. The name is a combination of the original table name and the partition name.

22.5.3 Example: Recovering a Table into a New Schema

This example recovers multiple tables into a new schema that is different from the source schema.

In this example, the HR.DEPARTMENTS and SH.CHANNELS tables need to be recovered to the state that they were in one day ago, before a logical corruption occurred. The recovered tables must be renamed as NEW_DEPARTMENTS and NEW_CHANNELS and imported into the EXAMPLE schema. The schema EXAMPLE exists at the time this example is run.

The REMAP TABLE clause is used to indicate how the source schema is mapped to a new target schema. The auxiliary destination used during the recovery process is /tmp/auxdest.

  1. Perform the planning tasks required to recover tables from RMAN backups. In this example, you need to recover tables to a point in time specified by an expression that uses SYSDATE.
  2. Start an RMAN session and connect to the target database as described in "Making Database Connections with RMAN".
  3. Recover the HR.DEPARTMENTS and SH.CHANNELS tables, rename them to NEW_DEPARTMENTS and NEW_CHANNELS respectively, and then import them into the EXAMPLE schema.

    The following RECOVER command performs the required table recovery: 

    RECOVER TABLE HR.DEPARTMENTS, SH.CHANNELS
UNTIL TIME 'SYSDATE – 1' 
AUXILIARY DESTINATION '/tmp/auxdest'
REMAP TABLE hr.departments:example.new_departments, sh.channels:example.new_channels;