The Top 12 Features new to Oracle 12c

The top 12 new features of Oracle 12c!
David Yahalom,
CTO, NAYA Technologies 
www.naya-tech.com
Email:
davidy@naya-tech.co.il

www.naya-tech.com | 5 Penn Plaza, 23rd floor Manhattan, New York 10001 +1.212.896.3945
Improved column defaults
SQL> create sequence s;
Sequence created.
SQL> create table my_table
2 ( x int
3 default s.nextval
4 primary key,
5 y varchar2(30)
6 );
Table created.
• > Sequences supported for columns
without a trigger!

• > We can now use an IDENTITY type!
•
• > Generates a sequence and associate that
sequence with the table.
create table my_Table
(x int generated as identity
primary key,
y varchar2(30));

create table t
(x int generated by default
as identity
(start with 42
increment by 1000 )
primary key,
y varchar2(30))
• > Complex identity values supported

Increased size limits
> VARCHARS can go up to 32K!
Set MAX_STRING_SIZE init.ora parameter to
EXTENDED.
Run @?/rdbms/admin/utl32k.sql
create table t ( x varchar(32767) );
>> Actually stored as LOB
>> In-row <= 4K, out of row > 4K…

Increased size limits
> But now you can use RPAD/LPAD/TRIM !
SQL> insert into my_tab values ( rpad('*',
32000,'*') );
1 row created.
SQL> select length(x) from my_tab;
LENGTH(X)
——————————————
32000
(previously string built-in functions would have been
able to return only 4,000 bytes)

Improved top-N queries
> New Row limiting clause for result set
pagination.
> Support for the ANSI-standard FETCH FIRST/
NEXT and OFFSET
create table t
as select * from all_objects;
create index t_idx on t(owner,object_name);

> Retrieve the first five rows after sorting by
OWNER and OBJECT_NAME
select owner, object_name, object_id
from t
order by owner, object_name
FETCH FIRST 5 ROWS ONLY;

> The optimizer is rewriting the query to use
analytics!
…
——————————————————————————————————————————————————————————————————————————————
| Id |Operation | Name|Rows |Bytes |Cost (%CPU)|Time |
——————————————————————————————————————————————————————————————————————————————
| 0|SELECT STATEMENT | | 5 | 1450 | 7 (0)|00:00:01|
|* 1| VIEW | | 5 | 1450 | 7 (0)|00:00:01|
|* 2| WINDOW NOSORT STOPKEY | | 5 | 180 | 7 (0)|00:00:01|
| 3| TABLE ACCESS BY INDEX ROWID|T |87310 | 3069K| 7 (0)|00:00:01|
| 4| INDEX FULL SCAN |T_IDX| 5 | | 3 (0)|00:00:01|
——————————————————————————————————————————————————————————————————————————————
Predicate Information (identified by operation id):
—————————————————————————————————————————————————————————————————
1 - filter("from$_subquery$_003"."rowlimit_$$_rownumber"<=5)
2 - filter(ROW_NUMBER() OVER ( ORDER BY "OWNER","OBJECT_NAME")<=5)

> To paginate through a result set: 
(Get N rows at a time from a specific page in the result set
—add the OFFSET clause).
select owner, object_name, object_id
from t
order by owner, object_name
OFFSET 5 ROWS FETCH NEXT 5 ROWS ONLY;

Temporary UNDO
> Previously:  
 
Temporary tablespace DML
Generates UNDO in the UNDO TBS 
(for read consistency)
UNDO TBS changes required REDO for crash
recovery

Temporary UNDO
Temp TBS
Redo logs
Undo TBS
Bulk Load

Temporary UNDO
Temp TBS & Temporary Undo
Redo logs
Undo TBS
Bulk Load
Permanent tables
Operations on temporary tables will
no longer generate redo.

Temporary UNDO
> Can be used with Active DataGuard!
Read-only replicated tables
Read / Write temporary table
(intermediate query results)
Source Database

Temporary UNDO
alter session
set temp_undo_enabled = true;
update my_table set object_name =
lower(object_name);
87310 rows updated.
Statistics
———————————————————————————————
…
0 redo size
…

New partitioning features
> Move a partition ONLINE!
(non-blocking DDL, allow DML)
alter table test_tbl move partition p1 ONLINE;

Transaction Guard
> For database developers.
> API that returns the outcome of the
last transaction.
> Provide protection for sensitive
transactions that are allowed to only
happen once.

Transaction Guard
> Without:

Transaction Guard
CallableStatement c = conn2.prepareCall(
"declare b1 boolean; b2 boolean; begin"
+"DBMS_APP_CONT.GET_LTXID_OUTCOME(?,b1,"
+"b2); ? := case when B1 then "
+"'COMMITTED' else 'UNCOMMITTED' end; "
+"end;");

Transaction Guard
> With:

Adaptive Execution Plans
> Before Oracle 12c, plans were fixed for the
first execution.
> Unexpected high row counts may make first plan
suboptimal.
> With 12, the Optimizer can now generate
plan + subplans. 
> Optimizer picks final plan based on cardinality
during first execution.
> “Changes its mind” in realtime!

--------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 23 | 4 (0)| 00:00:01 |
| 1 | HASH UNIQUE | | 1 | 23 | 4 (0)| 00:00:01 |
|- * 2 | HASH JOIN SEMI | | 1 | 23 | 4 (0)| 00:00:01 |
| 3 | NESTED LOOPS SEMI | | 1 | 23 | 4 (0)| 00:00:01 |
|- 4 | STATISTICS COLLECTOR | | | | | |
| * 5 | TABLE ACCESS FULL | DEPARTMENTS | 1 | 16 | 3 (0)| 00:00:01 |
| * 6 | TABLE ACCESS BY INDEX ROWID BATCHED| EMPLOYEES | 1 | 7 | 1 (0)| 00:00:01 |
| * 7 | INDEX RANGE SCAN | EMP_DEPARTMENT_IX | 10 | | 0 (0)| 00:00:01 |
|- * 8 | TABLE ACCESS FULL | EMPLOYEES | 1 | 7 | 1 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------------------------
Note
-----
- this is an adaptive plan (rows marked '-' are inactive)
> STATISTICS COLLECTOR buffers the rows and able
to switch to HASH JOIN when cardinality becomes
higher than what was estimated.

Rejected!
Accepted!

Enhanced Statistics
> New histograms: Top, Hybrid. 
> New Dynamic Sampling:  
Dynamic Sampled statistics (now
Dynamic Statistics) can be reused.
 
If defined at 2 (which is the default) dynamics statistics will be
gathered if at leat one table in the query has no statistics.
If defined to 11 the database will use dynamic statistics
automatically when statistics are missing, statistics are stale,
statistics are insufficient.

Enhanced Statistics
> Automatically compute statistics
during loads (CATS).

Data Optimisation and ILM
> Oracle 12c creates “Heat Maps”  
- tracks and marks data at the row and block
level as it goes through life cycle changes. 
 
> Automatic Data Optimization works with the
Heat Map feature and allows us to create
policies.
> Automatic Data Optimization allows you to
create policies for data compression and data
movement, to implement storage tiers.

> Data can be:
Hot: the object is actively in Read/Write.
Warm: the object which is accessed in
reads only
Cold: the object is not participating in
any kind of activity.

SQL> alter session set heat_map=on;
SQL> select * from scott.emp;
EMPNO ENAME JOB MGR HIREDATE SAL
COMM DEPTNO
---------- ---------- --------- ---------- --------- ----------
---------- ----------
7369 SMITH CLERK 7902 17-DEC-80 800
20
7499 ALLEN SALESMAN 7698 20-FEB-81 1600
…

select object_name, track_time "Tracking Time",
segment_write "Segment write",
full_scan "Full Scan",
lookup_scan "Lookup Scan"
from DBA_HEAT_MAP_SEG_HISTOGRAM
where object_name='MYOBJECTS'
and owner = 'SCOTT';
OBJECT_NAME
-------------------------------------------------------------
-------------------
Tracking Time Segment write Full Scan Lookup Scan
------------------ -------------- ------------ ------------
MYOBJECTS
09-sep-13 02:40:14 NO YES NO

ALTER TABLE scott.myobjects ILM ADD POLICY ROW
STORE
COMPRESS ADVANCED SEGMENT AFTER 30 DAYS OF NO
MODIFICATION;

Row Pattern Matching
> An extension to the SELECT
statement using MATCH_RECOGNIZE
that allows us to identify patterns
across sequences of rows.

Row Pattern Matching
PATTERN (STRT DOWN+ UP+)
DEFINE
DOWN AS
DOWN.price < PREV(DOWN.price),
UP AS UP.price > PREV(UP.price)
XYZ 13-MAR-15 35 ***********************************
XYZ 14-MAR-15 34 **********************************
XYZ 15-MAR-15 33 *********************************
XYZ 16-MAR-15 34 **********************************
XYZ 17-MAR-15 35 ***********************************
XYZ 18-MAR-15 36 ************************************
XYZ 19-MAR-15 37 *************************************
XYZ 20-MAR-15 36 ************************************
XYZ 21-MAR-15 35 ***********************************
XYZ 22-MAR-15 34 **********************************
XYZ 23-MAR-15 35 ***********************************
XYZ 24-MAR-15 36 ************************************
XYZ 25-MAR-15 37 *************************************
Any record, followed by one or more records in which the price of the stock goes
down, followed by one or more records in which the stock price increases.

PL/SQL enhancements.
> Define PL/SQL Subprograms in a
SQL Statement.
> Why would a developer want to copy
logic from a PL/SQL function into a
SQL statement?
To improve performance.
> No context switch to the PL/SQL
engine.

Pluggable Databases
A PDB is a self-contained, fully
functional Oracle Database, and
includes its own system, sysaux
and user tablespaces.

Pluggable Databases
> CDB: Similar to a conventional Oracle
database.
> Contains most of the working parts you will be already
familiar with (controlfiles, datafiles, undo, tempfiles, redo
logs etc.).
> Contains the data dictionary for those objects that are
owned by the root container and those that are visible to
all PDBs.

Pluggable Databases
> PDB: Contains information specific to itself.
> Made up of datafiles and tempfiles to handle it's own
objects: includes it's own data dictionary, containing
information about only those objects that are specific to
the PDB.

Pluggable Databases

Pluggable Databases
> Allows databases to be moved easily
> Allows quick patching and upgrading to future
versions.
A PDB can be unplugged from a 12.1 CBD and plugged
into a 12.2 CDB, effectively upgrading it in seconds.

Pluggable Databases
12.1.0.2

Thank You
Please visit us at www.naya-tech.com

The Top 12 Features new to Oracle 12c

Recommended

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to The Top 12 Features new to Oracle 12c (20)

Recently uploaded (20)

The Top 12 Features new to Oracle 12c