Oracle – How to Shrink Space on Table with Function-Based Index

The solution is dbms_redefinition package. Basically redefition will move, online, to a new table (called interim table), then move all dependent objets like indexes and last exchange the interim table with the original one.

What you have to do is:

• Create the interim table withe correct column definition
• Run redefinition
• Drop the old table

Use dbms_redefinition in the following way:

-- DETERMINE IF THE ORIGINAL TABLE CAN BE REDEFINED ONLINE
BEGIN
DBMS_REDEFINITION.CAN_REDEF_TABLE('SCHEMA','YOURTABLE', DBMS_REDEFINITION.CONS_USE_ROWID);   
END;
/

-- BEGIN THE REDEFINITION
BEGIN
DBMS_REDEFINITION.START_REDEF_TABLE(
UNAME => 'SCHEMA',
ORIG_TABLE => 'YOURTABLE',
INT_TABLE => 'INTERIM_YOURTABLE'
OPTIONS_FLAG => DBMS_REDEFINITION.CONS_USE_ROWID);
END;
/

-- IF THE TABLE HAD DEPENDENCIES (INDEXES ... CONSTRAINTS ... TRIGGERS)
-- THIS WOULD BE THE POINT AT WHICH THEY WOULD HAVE BEEN COPIED
DECLARE 
error_count pls_integer := 0;
BEGIN
DBMS_REDEFINITION.COPY_TABLE_DEPENDENTS('SCHEMA', 'YOURTABLE', 'INTERIM_YOURTABLE', 
dbms_redefinition.cons_orig_params, TRUE, TRUE, TRUE, FALSE, error_count);
DBMS_OUTPUT.PUT_LINE('errors := ' || TO_CHAR(error_count));
END;
/

-- FINISH THE REDEFINITION
exec DBMS_REDEFINITION.FINISH_REDEF_TABLE('SCHEMA','YOURTABLE','INTERIM_YOURTABLE');

If you run redefinition with SYSTEM user there is no problem. If you want to execute it with a less privileged user you have to trick some privileges in order to get it working. Privileges required are:

• Execute privilege to DBMS_REDEFINITION
• Create any table
• Alter any table
• Drop any table
• Lock any table
• Select any table

Tables with the following characteristics cannot be redefined online:

• [9.0.1]Tables with no primary keys
• Tables that have materialized view logs defined on them
• [9i] Tables that are materialized view container tables and AQ tables
• [10g] Tables that are replicated in an n-way master configuration can be redefined, but horizontal subsetting (subset of rows in the table), vertical subsetting (subset of columns in the table), and column transformations are not allowed.
• The overflow table of an IOT table
• Tables with fine-grained access control (row-level security)
• Tables with BFILE columns
• Tables with LONG columns can be redefined online, but those columns must be converted to CLOBS. Also, LONG RAW columns must be converted to BLOBS. Tables with LOB columns are acceptable.
• Tables in the SYS and SYSTEM schema
• Temporary tables

Other restrictions:

• A subset of rows in the table
• Only simple deterministic expressions, sequences, and SYSDATE can be used when mapping the columns in the interim table to those of the original table. For example, subqueries are not allowed.
• If new columns are being added with no column mappings, then they must not be declared NOT NULL until the redefinition is complete.
• There cannot be any referential constraints between the table being redefined and the interim table.
• Table redefinition cannot be done NOLOGGING.
• [10g] For materialized view logs and queue tables, online redefinition is restricted to changes in physical properties.
• You cannot convert a nested table to a VARRAY.

Here are two best practices for partitioning that pertain to the question:

1. Keep an empty staging partition at the leftmost and rightmost ends of the partition range to ensure that the partitions split when loading in new data, and merge, after unloading old data, do not cause data movement.
2. Do not split or merge a partition already populated with data because this can cause severe locking and explosive log growth.

http://www.informit.com/articles/article.aspx?p=1946159&seqNum=5

If the leftmost end of your partition is empty, use ALTER PARTITION FUNCTION SPLIT RANGE to add new ranges to the partition function.

To check if the leftmost partition is empty, use a query like the following:

DECLARE @PartitionFunctionName sysname = 'YourPartitionFunctionNameHere';

SELECT 
p.partition_number, SUM(pst.row_count) RowCountInPartition, pf.name PartitionFunction, ps.name PartitionScheme
FROM sys.dm_db_partition_stats pst
INNER JOIN sys.partitions p ON pst.partition_id = p.partition_id
INNER JOIN sys.indexes i ON p.object_id = i.object_id AND p.index_id = i.index_id
INNER JOIN sys.partition_schemes ps ON ps.data_space_id = i.data_space_id
INNER JOIN sys.partition_functions pf ON ps.function_id = pf.function_id
WHERE pf.name = @PartitionFunctionName
GROUP BY p.partition_number, pf.name, ps.name;

If the first partition is not empty, the best practices recommend that you create a new function with all values, create a new table on that function, then insert the data to the new table.

Also, if the left partition just has a few records, a split may be fine. Not sure on that as I've never tried it.

Whatever you do, make sure to leave some empty partitions at the leftmost and rightmost partition when you're finished. I might even go so far as to creating partition ranges for 0 and 1, then add a check constraint to prevent the first partition from getting data in it. Do the same thing for the end.