jump to navigation

12c Partial Indexes For Partitioned Tables Part II (Vanishing Act) July 12, 2013

Posted by Richard Foote in 12c, Local Indexes, Oracle Indexes, Partial Indexes, Partitioning.
2 comments

In Partial Indexes Part I, we looked at how it was possible with the 12c database  to create a Partial Index based on data from only selected table partitions. The resultant Partial Index can be either a Global or Local Index.

In Part I, we only really looked at Global Indexes, so let’s look at a Local Index example. Using the same Partitioned Table example as before:

SQL> create table pink_floyd (id number, status varchar2(6), name varchar2(30))
indexing off
partition by range (id)
(partition pf1 values less than (1000001),
partition pf2 values less than (2000001) indexing off,
partition pf3 values less than (maxvalue) indexing on);
Table created.

This time, we’ll create a Local Partial Index:

SQL> create index pink_floyd_status_i on pink_floyd(status)
local indexing partial;

Index created.

If we look at the details of the resultant Local Index:

SQL> select index_name, partition_name, num_rows, status, leaf_blocks from dba_ind_partitions where index_name = 'PINK_FLOYD_STATUS_I';

INDEX_NAME           PARTITION_NAME    NUM_ROWS STATUS   LEAF_BLOCKS
-------------------- --------------- ---------- -------- -----------
PINK_FLOYD_STATUS_I  PK1                      0 UNUSABLE           0
PINK_FLOYD_STATUS_I  PK2                      0 UNUSABLE           0
PINK_FLOYD_STATUS_I  PK3                1000000 USABLE          2513

We can see that for those table partitions with INDEXING OFF, the associated Local Indexes have simply been made UNUSABLE. Since Unusable Indexes consume no storage, there is effectively no corresponding index segment for these index partitions.

For the one and only PK3 table partition with INDEXING ON, its associated Local Index has been created as normal. So the end result is very similar to the previous Global Index example, only those rows from the table partitions with the INDEXING ON property are effectively being indexed.

There is one scenario in which the creation of a Partial Index is not permitted, that is in the creation of a Unique Index or a Non-Unique Index to police a Primary Key or Unique Key constraint. Some examples:

SQL> create unique index pink_floyd_id_i on pink_floyd(id)
indexing partial;
create unique index pink_floyd_id_i on pink_floyd(id) indexing partial
*
ERROR at line 1:

ORA-14226: unique index may not be PARTIAL

SQL> alter table pink_floyd add constraint pink_floyd_pk primary key(id)
using index (create index pink_floyd_id_i on pink_floyd(id) indexing partial);
alter table pink_floyd add constraint pink_floyd_pk primary key(id) using index
(create index pink_floyd_id_i on pink_floyd(id) indexing partial)
*
ERROR at line 1:
ORA-14196: Specified index cannot be used to enforce the constraint.

SQL> create index pink_floyd_id_i on pink_floyd(id) indexing partial;

Index created.

SQL> alter table pink_floyd add primary key(id);
alter table pink_floyd add primary key(id)
*
ERROR at line 1:

ORA-01408: such column list already indexed

It clearly doesn’t make sense to create a Partial Unique Index or on a Non-Unique Index policing a PK or Unique Key constraint as it would be impossible to use such an index to guarantee the required unique property. With missing index entries associated with non-indexed partitions, how can Oracle determine whether a value from new row already exists or not ? It can’t and hence Oracle doesn’t permit the creation of such a Partial Index.

Partial Indexes can potentially be extremely useful in reducing unnecessary storage requirements, reducing index maintenance overheads and in improving performance by reducing index block accesses.

But they’re only useful (possible) with Partitioned Tables.

I’ll next look at another cool index improvement introduced with the Oracle 12c Database that’s associated with Partitioning, Asynchronous Global Index Maintenance …

Local Index Issue With Partitioned PK and Unique Key Constraints December 20, 2007

Posted by Richard Foote in Constraints, Index Access Path, Local Indexes, Oracle Indexes, Partitioning, Performance Tuning, Unique Indexes.
11 comments

Nuno Souto (Noons) also asked a really interesting question on my Differences between Unique and Non-Unique Indexes blog entry (comment 4) that I thought it worthy of a separate blog entry to do the answer justice. The question was:

“Isn’t it still the case that unique indexes cannot be locally partitioned unless the partition key is part of the index key? Not sure if 11g removes this. If still so, that would weigh heavily in favour of non-unique indexing for PK on a table potentially requiring local index partitions.”

Simplistically, the answer to the first part is Yes it is still the case, even in 11g and the answer to the second part is No, it wouldn’t weigh heavily in favour of non-unique indexing for PK on a table requiring local index partitions. It wouldn’t actually be a consideration at all.

Let me explain why.

Firstly, there is a really really good reason why Oracle doesn’t allow us to create a Unique Index in which the Partition key is not part of a Local Index. It’s called protecting us from ourselves !!

Let’s start by mentioning constraints again.

Remember, the main reason we have indexes policing PK and Unique constraints is so that Oracle can very quickly and efficiently determine whether or not a new value already exists. Do a quick index look-up, is the value there, yes or no, allow the insert (or update), yes or no.

Just imagine for one moment what would happen if Oracle actually allowed us to create a Unique Local index in which the index didn’t include the partitioned column(s).

Lets say a table is Range Partitioned on column ‘A’ and we try and create a Unique Local index on just column ‘B’. Let’s assume we have (say) 500 table partitions meaning we must therefore have 500 local index partitions as well. When we insert a new value for our unique index for value B, it will attempt to do so in the corresponding local index partition as governed by the value A for the new row. However Oracle can’t just check this one index partition for uniqueness to ensure value of column B doesn’t already exist, Oracle would need to check all 500 index partitions because it would be possible for our new value of column B to potentially have previously been inserted into any of the other 499 partitions !!

Each and every insert into our partitioned table (partitioned by column A) therefore would require Oracle to check all (say)500 index partitions each and every time to check for duplicates of column B. Again, it’s important to understand that any given value of column B could potentially be in any of the 500 partitions, IF Oracle allowed us to create a Local Partitioned Index just on column B.

Checking all 500 index partitions looking for a specific value of column B would obviously be impractical, inefficient and totally un-scalable. Therefore Oracle doesn’t allow us to do this. It doesn’t allow us to create a Local index in which the indexed columns does’t include the partitioning columns as well.

This is actually a good thing.

If you want to create a Unique index in a partitioned table, you MUST either add all the partitioned columns and make it part of the LOCAL unique index (so that way each and every insert would only have to check the one local partition as this value is known now it’s part of the index) or you must create it as a GLOBAL index (in which again, Oracle only has to check the one index structure).

It actually makes a lot of sense to do this.

Moving onto the second part of the question. Let’s just use a Local Non-Unique index to police our PK constraints then.

Fortunately this isn’t allowed either for exactly the same reasons. You can’t create a Local Non-unique index to police a PK (or Unique) constraint if the Constraint does not also include the partitioned columns. Otherwise again, Oracle would need to check each and every index partition to determine whether the constraint has been violated or not.

If you attempt to use an existing Local Non-Unique index to police a PK or Unique constraint that does not contain the partitioned columns, you will get an error saying it can’t create the (by default Global index) because the useless Local Non-Unique index (from a policing the constraint point of view) already exists.

Again if you want to create a Non-Unique index to police a PK or Unique constraint you must either ensure the constraint includes all the partitioned columns in which case it can be Local or you must use a Global Non-Unique index.

In other words, the rules apply equally to both Unique and Non-Unique indexes.

So it’s not really a case of Oracle not allowing one to create a Local Unique index without including the partitioned columns (although that’s of course true) but really a case of Oracle not allowing a PK or Unique *constraint*  to be policed via *any* Local index (whether Unique or Non-Unique), unless the partitioned columns are also included.

Little demo to illustrate: Local Index Issue With Partitioned PK and Unique Key Constraints