Use the ALTER TABLE ADD PARTITION statement to add a new partition to the "high" end (the point after the last existing partition). To add a partition at the beginning or in the middle of a table, use the SPLIT PARTITION clause.

For example, consider the table, sales, which contains data for the current month in addition to the previous 12 months. On January 1, 1999, you add a partition for January, which is stored in tablespace tsx.

ALTER TABLE sales
      ADD PARTITION jan99 VALUES LESS THAN ( '01-FEB-1999' )
      TABLESPACE tsx;

Local and global indexes associated with the range-partitioned table remain usable.

Consequently, if the table contains data, then it may take some time to add a hash partition.

The following statements show two ways of adding a hash partition to table scubagear. Choosing the first statement adds a new hash partition whose partition name is system generated, and which is placed in the default tablespace. The second statement also adds a new hash partition, but that partition is explicitly named p_named and is created in tablespace gear5.

ALTER TABLE scubagear ADD PARTITION;

ALTER TABLE scubagear
      ADD PARTITION p_named TABLESPACE gear5;

Indexes may be marked UNUSABLE as explained in the following table:

The following statement illustrates how to add a new partition to a list-partitioned table. In this example, physical attributes and NOLOGGING are specified for the partition being added.

ALTER TABLE q1_sales_by_region 
   ADD PARTITION q1_nonmainland VALUES ('HI', 'PR')
      STORAGE (INITIAL 20K NEXT 20K) TABLESPACE tbs_3
      NOLOGGING;

Any value in the set of literal values that describe the partition being added must not exist in any of the other partitions of the table.

You cannot add a partition to a list-partitioned table that has a default partition, but you can split the default partition. By doing so, you effectively create a new partition defined by the values that you specify, and a second partition that remains the default partition.

Local and global indexes associated with the list-partitioned table remain usable.

However, exchanging a partition of an interval-partitioned table that has not been materialized in the data dictionary, meaning to have an explicit entry in the data dictionary beyond the interval definition, you must manually materialize the partition using the ALTER TABLE LOCK PARTITION command.

To change the interval for future partitions, use the SET INTERVAL clause of the ALTER TABLE statement. The SET INTERVAL clause converts existing interval partitions to range partitions, determines the high value of the defined range partitions, and automatically creates partitions of the specified interval as needed for data that is beyond that high value. As a side effect, an interval-partitioned table does not have the notation of MAXVALUES.

You also use the SET INTERVAL clause to migrate an existing range partitioned or range-* composite partitioned table into an interval or interval-* partitioned table. To disable the creation of future interval partitions, and effectively revert to a range-partitioned table, use an empty value in the SET INTERVAL clause. Created interval partitions are transformed into range partitions with their current high values.

To increase the interval for date ranges, you must ensure that you are at a relevant boundary for the new interval. For example, if the highest interval partition boundary in your daily interval partitioned table transactions is January 30, 2007 and you want to change to a monthly partition interval, then the following statement results in an error:

ALTER TABLE transactions SET INTERVAL (NUMTOYMINTERVAL(1,'MONTH');

ORA-14767: Cannot specify this interval with existing high bounds

You must create another daily partition with a high bound of February 1, 2007 to successfully change to a monthly interval:

LOCK TABLE transactions PARTITION FOR(TO_DATE('31-JAN-2007','dd-MON-yyyy') 
   IN SHARE MODE;

ALTER TABLE transactions SET INTERVAL (NUMTOYMINTERVAL(1,'MONTH');

The lower partitions of an interval-partitioned table are range partitions. You can split range partitions to add more partitions in the range portion of the interval-partitioned table.

To disable interval partitioning on the transactions table, use:

ALTER TABLE transactions SET INTERVAL ();

Specifically, when there is a local index defined on a table and you issue the ALTER TABLE statement to add a partition, a matching partition is also added to the local index. The database assigns names and default physical storage attributes to the new index partitions, but you can rename or alter them after the ADD PARTITION operation is complete.

You can effectively specify a new tablespace for an index partition in an ADD PARTITION operation by first modifying the default attributes for the index. For example, assume that a local index, q1_sales_by_region_locix, was created for list partitioned table q1_sales_by_region. If before adding the new partition q1_nonmainland, as shown in Adding a Partition to a List-Partitioned Table, you had issued the following statement, then the corresponding index partition would be created in tablespace tbs_4.

ALTER INDEX q1_sales_by_region_locix
   MODIFY DEFAULT ATTRIBUTES TABLESPACE tbs_4;

Otherwise, it would be necessary for you to use the following statement to move the index partition to tbs_4 after adding it:

ALTER INDEX q1_sales_by_region_locix 
   REBUILD PARTITION q1_nonmainland TABLESPACE tbs_4;
 

You can add a partition to a hash partitioned global index using the ADD PARTITION syntax of ALTER INDEX. The database adds hash partitions and populates them with index entries rehashed from an existing hash partition of the index, as determined by the hash function. The following statement adds a partition to the index hgidx shown in Creating a Hash Partitioned Global Index:

ALTER INDEX hgidx ADD PARTITION p5;

You cannot add a partition to a range-partitioned global index, because the highest partition always has a partition bound of MAXVALUE. To add a new highest partition, use the ALTER INDEX SPLIT PARTITION statement.

When adding multiple partitions, local and global index operations are the same as when adding a single partition. Adding multiple partitions and subpartitions is only supported for range, list, and system partitions and subpartitions.

You can add multiple range partitions that are listed in ascending order of their upper bound values to the high end (after the last existing partition) of a range-partitioned or composite range-partitioned table, provided the MAXVALUE partition is not defined. Similarly, you can add multiple list partitions to a table using new sets of partition values if the DEFAULT partition does not exist.

Multiple system partitions can be added using a single SQL statement by specifying the individual partitions. For example, the following SQL statement adds multiple partitions to the range-partitioned sales table created in Example 4-1:

ALTER TABLE sales ADD 
  PARTITION sales_q1_2007 VALUES LESS THAN (TO_DATE('01-APR-2007','dd-MON-yyyy')),
  PARTITION sales_q2_2007 VALUES LESS THAN (TO_DATE('01-JUL-2007','dd-MON-yyyy')),
  PARTITION sales_q3_2007 VALUES LESS THAN (TO_DATE('01-OCT-2007','dd-MON-yyyy')),
  PARTITION sales_q4_2007 VALUES LESS THAN (TO_DATE('01-JAN-2008','dd-MON-yyyy'))
;

You can use the BEFORE clause to add multiple new system partitions in relation to only one existing partition. The following SQL statements provide an example of adding multiple individual partitions using the BEFORE clause:

CREATE TABLE system_part_tab1 (number1 integer, number2 integer) 
PARTITION BY SYSTEM
( PARTITION p1,
  PARTITION p2,
  PARTITION p3,
  PARTITION p_last);

ALTER TABLE system_part_tab1 ADD 
  PARTITION p4,
  PARTITION p5,
  PARTITION p6
  BEFORE PARTITION p_last;

SELECT SUBSTR(TABLE_NAME,1,18) table_name, TABLESPACE_NAME, 
   SUBSTR(PARTITION_NAME,1,16) partition_name 
   FROM USER_TAB_PARTITIONS WHERE TABLE_NAME='SYSTEM_PART_TAB1';
TABLE_NAME         TABLESPACE_NAME                PARTITION_NAME
------------------ ------------------------------ ----------------
SYSTEM_PART_TAB1   USERS                          P_LAST
SYSTEM_PART_TAB1   USERS                          P6
SYSTEM_PART_TAB1   USERS                          P5
SYSTEM_PART_TAB1   USERS                          P4
SYSTEM_PART_TAB1   USERS                          P3
SYSTEM_PART_TAB1   USERS                          P2
SYSTEM_PART_TAB1   USERS                          P1

The following statement reduces by one the number of partitions in a table by coalescing a partition.

ALTER TABLE ouu1
     COALESCE PARTITION;

The following statement distributes the contents of a subpartition of partition us_locations into one or more remaining subpartitions (determined by the hash function) of the same partition. For an interval-partitioned table, you can only coalesce hash subpartitions of materialized range or interval partitions. Basically, this operation is the inverse of the MODIFY PARTITION ADD SUBPARTITION clause discussed in Adding a Subpartition to a *-Hash Partitioned Table.

ALTER TABLE diving MODIFY PARTITION us_locations
     COALESCE SUBPARTITION;

The database selects the partition to coalesce based on the requirements of the hash partition. The following statement reduces by one the number of partitions in the hgidx index, created in Creating a Hash Partitioned Global Index:

ALTER INDEX hgidx COALESCE PARTITION;

This operation drops the data for the interval only and leaves the interval definition in tact. If data is inserted in the interval just dropped, then the database again creates an interval partition.

You can also drop range partitions in an interval-partitioned table. The rules for dropping a range partition in an interval-partitioned table follow the rules for dropping a range partition in a range-partitioned table. If you drop a range partition in the middle of a set of range partitions, then the lower boundary for the next range partition shifts to the lower boundary of the range partition you just dropped. You cannot drop the highest range partition in the range-partitioned section of an interval-partitioned table.

The following example drops the September 2007 interval partition from the sales table. There are only local indexes so no indexes are invalidated.

ALTER TABLE sales DROP PARTITION FOR(TO_DATE('01-SEP-2007','dd-MON-yyyy'));

If a global index partition is empty, then you can explicitly drop it by issuing the ALTER INDEX DROP PARTITION statement. But, if a global index partition contains data, then dropping the partition causes the next highest partition to be marked UNUSABLE. For example, you would like to drop the index partition P1, and P2 is the next highest partition. You must issue the following statements:

ALTER INDEX npr DROP PARTITION P1;
ALTER INDEX npr REBUILD PARTITION P2;

For example, the following SQL statement drops multiple partitions from the range-partitioned table sales.

ALTER TABLE sales DROP PARTITION sales_q1_2008, sales_q2_2008,
     sales_q3_2008, sales_q4_2008;

You cannot drop all the partitions of a table. When dropping multiple partitions, local and global index operations are the same as when dropping a single partition.

The following is an example of exchanging range partitions with a nonpartitioned table.

The following example shows a partition exchange for the interval_sales table, interval-partitioned using monthly partitions as of January 1, 2007. This example shows how to add data for June 2007 to the table using partition exchange load. Assume there are only local indexes on the interval_sales table, and equivalent indexes have been created on the interval_sales_june_2007 table.

ALTER TABLE interval_sales
  EXCHANGE PARTITION FOR (TO_DATE('01-JUN-2007','dd-MON-yyyy'))
  WITH TABLE interval_sales_jun_2007
  INCLUDING INDEXES;

Note the use of the FOR syntax to identify a partition that was system-generated. You can determine the partition name by querying the *_TAB_PARTITIONS data dictionary view to display the system-generated partition name.

Example 4-32 shows a partition exchange load scenario for the range-partitioned orders table, and the reference partitioned order_items table. The data in the order_items_dec_2006 table only contains order item data for orders with an order_date in December 2006.

You must use the UPDATE GLOBAL INDEXES or UPDATE INDEXES on the exchange partition of the parent table in order for the primary key index to remain usable. Note also that you must create or enable the foreign key constraint on the order_items_dec_2006 table in order for the partition exchange on the reference-partitioned table to succeed.

The following example illustrates this concept for a range-hash partitioned table.

First, create a hash partitioned table:

CREATE TABLE t1 (i NUMBER, j NUMBER)
     PARTITION BY HASH(i)
       (PARTITION p1, PARTITION p2);

Populate the table, then create a range-hash partitioned table as follows:

CREATE TABLE t2 (i NUMBER, j NUMBER)
     PARTITION BY RANGE(j)
     SUBPARTITION BY HASH(i)
        (PARTITION p1 VALUES LESS THAN (10)
            (SUBPARTITION t2_pls1,
             SUBPARTITION t2_pls2),
         PARTITION p2 VALUES LESS THAN (20)
            (SUBPARTITION t2_p2s1,
             SUBPARTITION t2_p2s2)
         );

It is important that the partitioning key in table t1 equals the subpartitioning key in table t2.

To migrate the data in t1 to t2, and validate the rows, use the following statement:

ALTER TABLE t2 EXCHANGE PARTITION p1 WITH TABLE t1
     WITH VALIDATION;

The following example converts the subpartition q3_1999_s1 of table sales into the nonpartitioned table q3_1999. Local index partitions are exchanged with corresponding indexes on q3_1999.

ALTER TABLE sales EXCHANGE SUBPARTITION q3_1999_s1
      WITH TABLE q3_1999 INCLUDING INDEXES;

The semantics are the same as described previously in Exchanging a Hash Partitioned Table with a *-Hash Partition. The following example shows an exchange partition scenario for a list-list partitioned table.

CREATE TABLE customers_apac
( id            NUMBER
, name          VARCHAR2(50)
, email         VARCHAR2(100)
, region        VARCHAR2(4)
, credit_rating VARCHAR2(1)
)
PARTITION BY LIST (credit_rating)
( PARTITION poor VALUES ('P')
, PARTITION mediocre VALUES ('C')
, PARTITION good VALUES ('G')
, PARTITION excellent VALUES ('E')
);

Populate the table with APAC customers. Then create a list-list partitioned table:

CREATE TABLE customers
( id            NUMBER
, name          VARCHAR2(50)
, email         VARCHAR2(100)
, region        VARCHAR2(4)
, credit_rating VARCHAR2(1)
)
PARTITION BY LIST (region)
SUBPARTITION BY LIST (credit_rating)
SUBPARTITION TEMPLATE
( SUBPARTITION poor VALUES ('P')
, SUBPARTITION mediocre VALUES ('C')
, SUBPARTITION good VALUES ('G')
, SUBPARTITION excellent VALUES ('E')
)
(PARTITION americas VALUES ('AMER')
, PARTITION emea VALUES ('EMEA')
, PARTITION apac VALUES ('APAC')
);

It is important that the partitioning key in the customers_apac table matches the subpartitioning key in the customers table.

Next, exchange the apac partition.

ALTER TABLE customers
EXCHANGE PARTITION apac
WITH TABLE customers_apac
WITH VALIDATION;

The semantics of the ALTER TABLE EXCHANGE PARTITION statement are the same as described previously in Exchanging a Hash Partitioned Table with a *-Hash Partition. The example below shows the orders table, which is interval partitioned by order_date, and subpartitioned by range on order_total. The example shows how to exchange a single monthly interval with a range-partitioned table.

CREATE TABLE orders_mar_2007
 ( id          NUMBER
  , cust_id     NUMBER
  , order_date  DATE
  , order_total NUMBER
 )
PARTITION BY RANGE (order_total)
 ( PARTITION p_small VALUES LESS THAN (1000)
  , PARTITION p_medium VALUES LESS THAN (10000)
  , PARTITION p_large VALUES LESS THAN (100000)
  , PARTITION p_extraordinary VALUES LESS THAN (MAXVALUE)
 );

Populate the table with orders for March 2007. Then create an interval-range partitioned table:

CREATE TABLE orders
 ( id          NUMBER
  , cust_id     NUMBER
  , order_date  DATE
  , order_total NUMBER
 )
PARTITION BY RANGE (order_date) INTERVAL (NUMTOYMINTERVAL(1,'MONTH'))
  SUBPARTITION BY RANGE (order_total)
  SUBPARTITION TEMPLATE
  ( SUBPARTITION p_small VALUES LESS THAN (1000)
   , SUBPARTITION p_medium VALUES LESS THAN (10000)
   , SUBPARTITION p_large VALUES LESS THAN (100000)
   , SUBPARTITION p_extraordinary VALUES LESS THAN (MAXVALUE)
  )
 (PARTITION p_before_2007 VALUES LESS THAN (TO_DATE('01-JAN-2007','dd-MON-yyyy')));

It is important that the partitioning key in the orders_mar_2007 table matches the subpartitioning key in the orders table.

Next, exchange the partition.

ALTER TABLE orders
  EXCHANGE PARTITION
   FOR (TO_DATE('01-MAR-2007','dd-MON-yyyy'))
   WITH TABLE orders_mar_2007
   WITH VALIDATION;

Nonadjacent range partitions cannot be merged. The resulting partition inherits the higher upper bound of the two merged partitions.

One reason for merging range partitions is to keep historical data online in larger partitions. For example, you can have daily partitions, with the oldest partition rolled up into weekly partitions, which can then be rolled up into monthly partitions, and so on.

Example 4-33 shows an example of merging range partitions using the ONLINE keyword.

Nonadjacent interval partitions cannot be merged. The first interval partition can also be merged with the highest range partition. The resulting partition inherits the higher upper bound of the two merged partitions.

Merging interval partitions always results in the transition point being moved to the higher upper bound of the two merged partitions. This result is that the range section of the interval-partitioned table is extended to the upper bound of the two merged partitions. Any materialized interval partitions with boundaries lower than the newly merged partition are automatically converted into range partitions, with their upper boundaries defined by the upper boundaries of their intervals.

For example, consider the following interval-partitioned table transactions:

CREATE TABLE transactions
( id               NUMBER
, transaction_date DATE
, value            NUMBER
)
PARTITION BY RANGE (transaction_date)
INTERVAL (NUMTODSINTERVAL(1,'DAY'))
( PARTITION p_before_2007 VALUES LESS THAN (TO_DATE('01-JAN-2007','dd-MON-yyyy')));

Inserting data into the interval section of the table creates the interval partitions for these days. The data for January 15, 2007 and January 16, 2007 are stored in adjacent interval partitions.

INSERT INTO transactions VALUES (1,TO_DATE('15-JAN-2007','dd-MON-yyyy'),100);
INSERT INTO transactions VALUES (2,TO_DATE('16-JAN-2007','dd-MON-yyyy'),600); 
INSERT INTO transactions VALUES (3,TO_DATE('30-JAN-2007','dd-MON-yyyy'),200);

Next, merge the two adjacent interval partitions. The new partition again has a system-generated name.

ALTER TABLE transactions
MERGE PARTITIONS FOR(TO_DATE('15-JAN-2007','dd-MON-yyyy'))
, FOR(TO_DATE('16-JAN-2007','dd-MON-yyyy'));

The transition point for the transactions table has now moved to January 17, 2007. The range section of the interval-partitioned table contains two range partitions: values less than January 1, 2007 and values less than January 17, 2007. Values greater than January 17, 2007 fall in the interval portion of the interval-partitioned table.

They do not need to be adjacent, as for range partitions, because list partitioning does not assume any order for partitions. The resulting partition consists of all of the data from the original two partitions. If you merge a default list partition with any other partition, then the resulting partition is the default partition.

The following statement merges two partitions of a table partitioned using the list method into a partition that inherits all of its attributes from the table-level default attributes. MAXEXTENTS is specified in the statement.

ALTER TABLE q1_sales_by_region 
   MERGE PARTITIONS q1_northcentral, q1_southcentral 
   INTO PARTITION q1_central 
      STORAGE(MAXEXTENTS 20);

The value lists for the two original partitions were specified as:

PARTITION q1_northcentral VALUES ('SD','WI')
PARTITION q1_southcentral VALUES ('OK','TX')

The resulting sales_west partition value list comprises the set that represents the union of these two partition value lists, or specifically:

('SD','WI','OK','TX')

The inheritance of properties is different when a *-hash partition is split, as opposed to when two *-hash partitions are merged. When a partition is split, the new partitions can inherit properties from the original partition because there is only one parent. However, when partitions are merged, properties must be inherited from the table level.

For interval-hash partitioned tables, you can only merge two adjacent interval partitions, or the highest range partition with the first interval partition. The transition point moves when you merge intervals in an interval-hash partitioned table.

The following example merges two range-hash partitions:

ALTER TABLE all_seasons
   MERGE PARTITIONS quarter_1, quarter_2 INTO PARTITION quarter_2
   SUBPARTITIONS 8;

The MERGE PARTITIONS and MERGE SUBPARTITIONS clauses are synonymous with the MERGE PARTITION and MERGE SUBPARTITION clauses.

For example, the following SQL statement merges four partitions into one partition and drops the four partitions that were merged.

ALTER TABLE t1 MERGE PARTITIONS p01, p02, p03, p04 INTO p0;

When merging multiple range partitions, the partitions must be adjacent and specified in the ascending order of their partition bound values. The new partition inherits the partition upper bound of the highest of the original partitions.

You can specify the lowest and the highest partitions to be merged when merging multiple range partitions with the TO syntax. All partitions between specified partitions, including those specified, are merged into the target partition. You cannot use this syntax for list and system partitions.

For example, the following SQL statements merges partitions p01 through p04 into the partition p0.

ALTER TABLE t1 MERGE PARTITIONS p01 TO p04 INTO p0;

List partitions and system partitions that you want to merge do not need to be adjacent, because no ordering of the partitions is assumed. When merging multiple list partitions, the resulting partition value list are the union of the set of partition value list of all of the partitions to be merged. A DEFAULT list partition merged with other list partitions results in a DEFAULT partition.

When merging multiple partitions of a composite partitioned table, the resulting new partition inherits the subpartition descriptions from the subpartition template, if one exists. If no subpartition template exists, then Oracle creates one MAXVALUE subpartition from range subpartitions or one DEFAULT subpartition from list subpartitions for the new partition. When merging multiple subpartitions of a composite partitioned table, the subpartitions to be merged must belong to the same partition.

When merging multiple partitions, local and global index operations and semantics for inheritance of unspecified physical attributes are the same for merging two partitions.

In the following SQL statement, four partitions of the partitioned by range table sales are merged. These four partitions that correspond to the four quarters of the oldest year are merged into a single partition containing the entire sales data of the year.

ALTER TABLE sales
  MERGE PARTITIONS sales_q1_2009, sales_q2_2009, sales_q3_2009, sales_q4_2009
  INTO PARTITION sales_2009;

The previous SQL statement can be rewritten as the following SQL statement to obtain the same result.

ALTER TABLE sales
  MERGE PARTITIONS sales_q1_2009 TO sales_q4_2009
  INTO PARTITION sales_2009;

For example, to move the most active partition to a tablespace that resides on its own set of disks (to balance I/O), not log the action, and compress the data, issue the following statement:

ALTER TABLE parts MOVE PARTITION depot2
     TABLESPACE ts094 NOLOGGING COMPRESS;

This statement always drops the old partition segment and creates a new segment, even if you do not specify a new tablespace.

If you are moving a partition of a partitioned index-organized table, then you can specify the MAPPING TABLE clause as part of the MOVE PARTITION clause, and the mapping table partition are moved to the new location along with the table partition.

For an interval or interval-* partitioned table, you can only move range partitions or interval partitions that have been materialized. A partition move operation does not move the transition point in an interval or interval-* partitioned table.

You can move a partition in a reference-partitioned table independent of the partition in the master table.

The following statement shows how to move data in a subpartition of a table. In this example, a PARALLEL clause has also been specified.

ALTER TABLE scuba_gear MOVE SUBPARTITION bcd_types 
     TABLESPACE tbs23 PARALLEL (DEGREE 2);

You can move a subpartition in a reference-partitioned table independent of the subpartition in the master table.

You can rebuild the entire index by rebuilding each partition individually using the ALTER INDEX REBUILD PARTITION statement. You can perform these rebuilds concurrently.

You can also simply drop the index and re-create it.

For example:

ALTER TABLE scubagear RENAME PARTITION sys_p636 TO tanks;

In this case, you would use the ALTER TABLE RENAME SUBPARTITION syntax.

In the SQL statement, you must specify values of the partitioning key column within the range of the partition at which to split the partition.

You can optionally specify new attributes for the partitions resulting from the split. If there are local indexes defined on the table, this statement also splits the matching partition in each local index.

If you do not specify new partition names, then the database assigns names of the form SYS_Pn. You can examine the data dictionary to locate the names assigned to the new local index partitions. You may want to rename them. Any attributes that you do not specify are inherited from the original partition.

The SPLIT PARTITION clause enables you to specify a list of literal values that define a partition into which rows with corresponding partitioning key values are inserted. The remaining rows of the original partition are inserted into a second partition whose value list contains the remaining values from the original partition. You can optionally specify new attributes for the two partitions that result from the split.

The following statement splits the partition region_east into two partitions:

ALTER TABLE sales_by_region 
   SPLIT PARTITION region_east VALUES ('CT', 'MA', 'MD') 
   INTO 
    ( PARTITION region_east_1 
         TABLESPACE tbs2,
      PARTITION region_east_2
        STORAGE (INITIAL 8M)) 
   PARALLEL 5;

The literal value list for the original region_east partition was specified as:

PARTITION region_east VALUES ('MA','NY','CT','NH','ME','MD','VA','PA','NJ')

The two new partitions are:

• region_east_1 with a literal value list of ('CT','MA','MD')

• region_east_2 inheriting the remaining literal value list of ('NY','NH','ME','VA','PA','NJ')

The individual partitions have new physical attributes specified at the partition level. The operation is executed with parallelism of degree 5.

You can split a default list partition just like you split any other list partition. This is also the only means of adding a new partition to a list-partitioned table that contains a default partition. When you split the default partition, you create a new partition defined by the values that you specify, and a second partition that remains the default partition.

Splitting a range partition in the interval-partitioned table is described in Splitting a Partition of a Range-Partitioned Table.

To split a materialized interval partition, you specify a value of the partitioning key column within the interval partition at which to split the partition. The first of the resulting two new partitions includes all rows in the original partition whose partitioning key column values map lower than the specified value. The second partition contains all rows whose partitioning key column values map greater than or equal to the specified value. The split partition operation moves the transition point up to the higher boundary of the partition you just split, and all materialized interval partitions lower than the newly split partitions are implicitly converted into range partitions, with their upper boundaries defined by the upper boundaries of the intervals.

You can optionally specify new attributes for the two range partitions resulting from the split. If there are local indexes defined on the table, then this statement also splits the matching partition in each local index. You cannot split interval partitions that have not yet been created.

The following example shows splitting the May 2007 partition in the monthly interval partitioned table transactions.

ALTER TABLE transactions
    SPLIT PARTITION FOR(TO_DATE('01-MAY-2007','dd-MON-yyyy'))
    AT (TO_DATE('15-MAY-2007','dd-MON-yyyy'));

This is the opposite of merging *-hash partitions. When you split *-hash partitions, the new subpartitions are rehashed into either the number of subpartitions specified in a SUBPARTITIONS or SUBPARTITION clause. Or, if no such clause is included, the new partitions inherit the number of subpartitions (and tablespaces) from the partition being split.

The inheritance of properties is different when a *-hash partition is split, versus when two *-hash partitions are merged. When a partition is split, the new partitions can inherit properties from the original partition because there is only one parent. However, when partitions are merged, properties must be inherited from table level defaults because there are two parents and the new partition cannot inherit from either at the expense of the other.

The following example splits a range-hash partition:

ALTER TABLE all_seasons SPLIT PARTITION quarter_1 
     AT (TO_DATE('16-dec-1997','dd-mon-yyyy'))
     INTO (PARTITION q1_1997_1 SUBPARTITIONS 4 STORE IN (ts1,ts3),
           PARTITION q1_1997­_2);

The rules for splitting an interval-hash partitioned table follow the rules for splitting an interval-partitioned table. As described in Splitting a Partition of an Interval-Partitioned Table, the transition point is changed to the higher boundary of the split partition.

• Splitting a *-List Partition

• Splitting a *-List Subpartition

Splitting a partition of a *-range partitioned table is similar to the description in Splitting a Partition of a Range-Partitioned Table. No subpartition range values can be specified for either of the new partitions. The new partitions inherit the subpartition descriptions from the original partition being split.

The following example splits the May 2007 interval partition of the interval-range partitioned orders table:

ALTER TABLE orders
    SPLIT PARTITION FOR(TO_DATE('01-MAY-2007','dd-MON-yyyy'))
    AT (TO_DATE('15-MAY-2007','dd-MON-yyyy'))
    INTO (PARTITION p_fh_may07,PARTITION p_sh_may2007);

This operation splits the interval partition FOR('01-MAY-2007') into two resulting partitions: p_fh_may07 and p_sh_may_2007. Both partitions inherit their subpartition descriptions from the original partition. Any interval partitions before the June 2007 partition have been converted into range partitions, as described in Merging Interval Partitions.

The ALTER TABLE SPLIT PARTITION statement provides no means of specifically naming subpartitions resulting from the split of a partition in a composite partitioned table. However, for those subpartitions in the parent partition with names of the form partition name_subpartition name, the database generates corresponding names in the newly created subpartitions using the new partition names. All other subpartitions are assigned system generated names of the form SYS_SUBPn. System generated names are also assigned for the subpartitions of any partition resulting from the split for which a name is not specified. Unnamed partitions are assigned a system generated partition name of the form SYS_Pn.

The following query displays the subpartition names and high values resulting from the previous split partition operation on table orders. It also reflects the results of other operations performed on this table in preceding sections of this chapter since its creation.

BREAK ON partition_name

SELECT partition_name, subpartition_name, high_value
FROM user_tab_subpartitions
WHERE table_name = 'ORCERS'
ORDER BY partition_name, subpartition_position;

PARTITION_NAME            SUBPARTITION_NAME              HIGH_VALUE
------------------------- ------------------------------ ---------------
P_BEFORE_2007             P_BEFORE_2007_P_SMALL          1000
                          P_BEFORE_2007_P_MEDIUM         10000
                          P_BEFORE_2007_P_LARGE          100000
                          P_BEFORE_2007_P_EXTRAORDINARY  MAXVALUE
P_FH_MAY07                SYS_SUBP2985                   1000
                          SYS_SUBP2986                   10000
                          SYS_SUBP2987                   100000
                          SYS_SUBP2988                   MAXVALUE
P_PRE_MAY_2007            P_PRE_MAY_2007_P_SMALL         1000
                          P_PRE_MAY_2007_P_MEDIUM        10000
                          P_PRE_MAY_2007_P_LARGE         100000
                          P_PRE_MAY_2007_P_EXTRAORDINARY MAXVALUE
P_SH_MAY2007              SYS_SUBP2989                   1000
                          SYS_SUBP2990                   10000
                          SYS_SUBP2991                   100000
                          SYS_SUBP2992                   MAXVALUE

However, you can split a global index partition as is done in the following example:

ALTER INDEX quon1 SPLIT 
    PARTITION canada AT ( 100 ) INTO 
    PARTITION canada1 ..., PARTITION canada2 ...);
ALTER INDEX quon1 REBUILD PARTITION canada1;
ALTER INDEX quon1 REBUILD PARTITION canada2;

The index being split can contain index data, and the resulting partitions do not require rebuilding, unless the original partition was previously marked UNUSABLE.

When splitting multiple partitions, the segment associated with the current partition is discarded. Each new partitions obtains a new segment and inherits all unspecified physical attributes from the current source partition.You can also use fast split when splitting into multiple partitions.

You can use the extended split syntax to specify a list of new partition descriptions similar to the create partitioned table SQL statements, rather than specifying the AT or VALUES clause. Additionally, the range or list values clause for the last new partition description is derived based on the high bound of the source partition and the bound values specified for the first (N-1) new partitions resulting from the split.

The following SQL statements are examples of splitting a partition into multiple partitions.

ALTER TABLE SPLIT PARTITION p0 INTO 
  (PARTITION p01 VALUES LESS THAN (25),
   PARTITION p02 VALUES LESS THAN (50), 
   PARTITION p03 VALUES LESS THAN (75),
   PARTITION p04);

ALTER TABLE SPLIT PARTITION p0 INTO 
  (PARTITION p01 VALUES LESS THAN (25),
   PARTITION p02);

In the second SQL example, partition p02 has the high bound of the original partition p0.

To split a range partition into N partitions, (N-1) values of the partitioning key column must be specified within the range of the partition at which to split the partition. The new non-inclusive upper bound values specified must be in ascending order. The high bound of Nth new partition is assigned the value of the high bound of the partition being split. The names and physical attributes of the N new partitions resulting from the split can be optionally specified.

To split a list partition into N partitions, (N-1) lists of literal values must be specified, each of which defines the first (N-1) partitions into which rows with corresponding partitioning key values are inserted. The remaining rows of the original partition are inserted into the Nth new partition whose value list contains the remaining literal values from the original partition. No two value lists can contain the same partition value. The (N-1) value lists that are specified cannot contain all of the partition values of the current partition because the Nth new partition would be empty. Also, the (N-1) value lists cannot contain any partition values that do not exist for the current partition.

When splitting a DEFAULT list partition or a MAXVALUE range partition into multiple partitions, the first (N-1) new partitions are created using the literal value lists or high bound values specified, while the Nth new partition resulting from the split have the DEFAULT value or MAXVALUE. Splitting a partition of a composite partitioned table into multiple partitions assumes the existing behavior with respect to inheritance of the number, names, bounds and physical properties of the subpartitions of the new partitions resulting from the split. The SPLIT_TABLE_SUBPARTITION clause is extended similarly to allow split of a range or list subpartition into N new subpartitions.

The behavior of the SQL statement with respect to local and global indexes remains unchanged. Corresponding local index partition are split into multiple partitions. If the partitioned table contains LOB columns, then existing semantics for the SPLIT PARTITION clause apply with the extended syntax; that is, LOB data and index segments is dropped for current partition and new segments are created for each LOB column for each new partition. Fast split optimization is applied to multipartition split operations when required conditions are met.

For example, the following SQL statement splits the sales_Q4_2007 partition of the partitioned by range table sales splits into five partitions corresponding to the quarters of the next year. In this example, the partition sales_Q4_2008 implicitly becomes the high bound of the split partition.

ALTER TABLE sales SPLIT PARTITION sales_Q4_2007 INTO
( PARTITION sales_Q4_2007 VALUES LESS THAN (TO_DATE('01-JAN-2008','dd-MON-yyyy')),
  PARTITION sales_Q1_2008 VALUES LESS THAN (TO_DATE('01-APR-2008','dd-MON-yyyy')),
  PARTITION sales_Q2_2008 VALUES LESS THAN (TO_DATE('01-JUL-2008','dd-MON-yyyy')),
  PARTITION sales_Q3_2008 VALUES LESS THAN (TO_DATE('01-OCT-2008','dd-MON-yyyy')),
  PARTITION sales_Q4_2008);

For the sample table customers partitioned by list, the following statement splits the partition Europe into three partitions.

ALTER TABLE list_customers SPLIT PARTITION Europe INTO
  (PARTITION western-europe VALUES ('GERMANY', 'FRANCE'),
   PARTITION southern-europe VALUES ('ITALY'), 
   PARTITION rest-europe);

The corresponding partitions of local indexes are truncated in the operation. Global indexes must be rebuilt unless UPDATE INDEXES is specified.

In the following example, the ALTER TABLE SQL statement truncates multiple partitions in a table. Note that the data is truncated, but the partitions are not dropped.

When truncating a subpartition, corresponding local index subpartitions are also truncated.

In the following example, the ALTER TABLE statement truncates data in subpartitions of a table. In this example, the space occupied by the deleted rows is made available for use by other schema objects in the tablespace with the DROP STORAGE clause. Note that the data is truncated, but the subpartitions are not dropped.

When the CASCADE option is specified for TRUNCATE TABLE, the truncate table operation also truncates child tables that reference the targeted table through an enabled referential constraint that has ON DELETE CASCADE enabled. This cascading action applies recursively to grandchildren, great-grandchildren, and so on. After determining the set of tables to be truncated based on the enabled ON DELETE CASCADE referential constraints, an error is raised if any table in this set is referenced through an enabled constraint from a child outside of the set. If a parent and child are connected by multiple referential constraints, a TRUNCATE TABLE CASCADE operation targeting the parent succeeds if at least one constraint has ON DELETE CASCADE enabled.

Privileges are required on all tables affected by the operation. Any other options specified for the operation, such as DROP STORAGE or PURGE MATERIALIZED VIEW LOG, apply for all tables affected by the operation.

When the CASCADE option is specified, the TRUNCATE PARTITION and TRUNCATE SUBPARTITION operations cascade to reference partitioned tables that are children of the targeted table. The TRUNCATE can be targeted at any level in a reference partitioned hierarchy and cascades to child tables starting from the targeted table. Privileges are not required on the child tables, but the usual restrictions on the TRUNCATE operation, such as the table cannot be referenced by an enabled referential constraint that is not a partitioning constraint, apply for all tables affected by the operation.

The CASCADE option is ignored if it is specified for a table that does not have reference partitioned children. Any other options specified for the operation, such as DROP STORAGE or UPDATE INDEXES, apply to all tables affected by the operation.

The cascade options are off by default so they do not affect Oracle Database compatibility.

ALTER TABLE sales 
     TRUNCATE PARTITION dec2016
     DROP STORAGE
     CASCADE
     UPDATE INDEXES;