What table column size is needed to accomodate Unicode characters

Hi guys,
I have encounter something which i dont understand and i hope gurus here will shed some light on me.
I am running a non-unicode database and i decided to port the data over to a unicode database.
So
1) i export the schema out --> data.dmp
2) then i create the unicode database + create a user
3) then i import the schema into the database
during the imp i can see that character conversion will take place.
During importing of data into the unicode database
I encounter some error
saying column size is too small
so i went to check the row that has the column value that is too large to fit in the table.
I realise it has some [][][][] data.. so i went to the live non-unicode database and find the row. Indeed it has some [][][][] rubbish data which i feel that someone has inserted other language then english into the database.
But regardless,
I went to modify the column size to a larger size, now the row can be accommodated. However the data is still [][][].
q1) why so ? since now my database is unicode, during the import, this column data [][][] should be converted to unicode already but i still have problem seeing what language it is.
q2) why at the non-unicode database, the [][][] data can fit into the table column size, but on unicode database, the same table column size need to be increase ?
q3) while doing more research on unicode, it was said that unicode character takes up 2 byte per character. Alot of my table data are exactly the same size of the table column size.
E.g Name VARCHAR2(5);
value - 'Peter'
Now if converting to unicode, characters will take 2byte instead of 1, isnt 'PETER' going to take up 10byte ( 2 byte per character ),
why is it that i can still accomodate the data into the table column ?
q4) now with unicode database up, i will be supporting different language characters around the world. How big should i set my column size to ? the longest a name can get ? or ?
Thanks guys!

/// does oracle automatically "look" at the each and individual characters in a word and determine how much byte it should take.
Characters usually originate from a keyboard, which has an associated keyboard layout and an associated character set encoding (a.k.a code page, a.k.a. encoding). This means, the keyboard driver knows that when a key with a letter "á" on it is pressed on a French keyboard, and the associated character set encoding is MS Code Page 1252 (Oracle name WE8MSWIN1252), then one byte with the value 225 is generated. If the associated character set encoding is UTF-16LE (standard internal Windows encoding), two bytes 225 and 0 are generated. When the generated bytes travel through APIs, they may undergo character set conversions from one encoding to another encoding. The conversion algorithms use translation tables to find out how to translate given byte sequence from one encoding to another encoding. In case of translation from WE8MSWIN1252 to AL32UTF8, Oracle will know that the byte sequence resulting from conversion of the code 225 should be 195 followed by 161. For a Chinese characters, for example when converting it from ZHS16GBK, Oracle knows the resulting sequence as well, and this sequence is usually 3 bytes.
This is how AL32UTF8 data gets into a database. Now, when Oracle processes a multibyte string, and needs to look at individual characters, for example to count them with LENGTH, or take a substring with SUBSTR, it uses information it has about the structure of the character set. Multibyte character sets are of two type: fixed-width and variable-width. Currently, Oracle supports only one fixed-width multibyte character set in the database: AL16UTF16, which is Oracle's name for Unicode UTF-16BE encoding. It supports this character set for NCHAR/NVARCHAR2/NCLOB data types only. This character set uses two bytes per each character code. To find the next code, 2 is simply added to the string pointer.
All other Oracle multibyte character sets are variable-width character sets, including AL32UTF8. In most cases, the length of each character code can be determined by looking at its first byte. In AL32UTF8, the number of 1-bits in the most significant positions in the first byte before the first 0-bit tells how many bytes a character has. 0 such bits means 1 byte (such codes are identical to 7-bit ASCII), 2 such bits mean two bytes, 3 bits mean 3 bytes, 4 bits mean four bytes. 1 bit (e.g. the bit sequence 10) starts each second, third or fourth byte of a code.
In other ASCII-based multibyte character sets, the number of bytes is usually determined by the value range of the first byte. Bytes below 128 means a one-byte code, bytes above 128 begin a two- or three-byte sequence, depending on the range.
There are also EBCDIC-based (mainframe) multibyte character sets, a.k.a shift-sensitive character sets, where a sequence of two-byte codes is introduced by inserting the SO character (code 14=0x0e) and ended by inserting the SI character (code 15=0x0f). There are also character sets, like ISO-2022-JP, which use more complicated byte sequences to define the length and meaning of byte sequences but Oracle supports them only in limited number of places.
/// e.g i have a word with 4 character. the 3rd character will be a chinese character..the rest are ascii character
/// will oracle use 4 byte per character regardless its ascii(english) or chinese
No.
/// or it will use 1 byte per english character then 3 byte for the chinese character ? e.g.total - 6 bytes taken
It will use 6 bytes.
Thnx,
Sergiusz

Similar Messages

Changing column size

Hi
I am working on oracle 8i. I have requirement to change a column size from varchar(3) to varchar2(4). I know the command to change the column size. all I need to know, their are 6 child tables. they are referencing this tables column.
what is the best way do I need to change child tables column size first and then i need to change column sixe of main table. do i need to consider to change views and triggers ans indexex.
please help me

Hello,
You would increase dramatically your chances to get an answer by posting in the Database forum.
Francois

I need the "Real" Table / Table Column names from a sql view

Firstly, we have a system with ~1000 tables and ~250 views. We have field level security on the table columns (in our code)
Now we have a lot of views which select data from the tables, but I need to get the "Table Column Name" that is linked in the view.
I know there are view columns that is not linked to a specific table column, (or concatenation of columns) so those columns can just return NULL.
From the sample you will see there is a view selecting data from another view. I know it is not the best sql performance (execution path) but that is another topic on its own.
I went through a lot of the sys.* stored procs or sys.* views to try and figure out if there is a view that we can use to get the expected results.
here is the sql code:
if
Exists (select
* from
sys.all_objects
where name =
'AliasView2')
drop view dbo.AliasView2
if
Exists (select
* from
sys.all_objects
where name =
'AliasView1')
drop view dbo.AliasView1
if
Exists (select
* from
sys.all_objects
where name =
'Table4')
BEGIN
alter table dbo.Table4
DROP CONSTRAINT [FK_T4_T3]
alter table dbo.Table4
DROP CONSTRAINT [PK_T4_Constraint]
drop table dbo.Table4
END
if
Exists (select
* from
sys.all_objects
where name =
'Table3')
BEGIN
alter table dbo.Table3
DROP CONSTRAINT [FK_T3_T2]
alter table dbo.Table3
DROP CONSTRAINT [PK_T3_Constraint]
drop table dbo.Table3
END
if
Exists (select
* from
sys.all_objects
where name =
'Table2')
BEGIN
alter table dbo.Table2
DROP CONSTRAINT [FK_T2_T1]
alter table dbo.Table2
DROP CONSTRAINT [PK_T2_Constraint]
drop table dbo.Table2
END
if
Exists (select
* from
sys.all_objects
where name =
'Table1')
BEGIN
alter table dbo.Table1
DROP CONSTRAINT [PK_T1_Constraint]
drop table dbo.Table1
END
create
Table dbo.Table1
T1_PK        int
NOT NULL
Identity(1, 1)
CONSTRAINT [PK_T1_Constraint]
PRIMARY KEY (T1_PK),
T1_Field1    varchar
NULL,
T1_Field2    varchar
NULL,
create
Table dbo.Table2
T2_PK        int
NOT NULL
Identity(1, 1)
CONSTRAINT [PK_T2_Constraint]
PRIMARY KEY (T2_PK),
T2_Field1    varchar
NULL,
T2_Field2    varchar
NULL,
T2_FK        int
NOT NULL
CONSTRAINT [FK_T2_T1]
FOREIGN KEY (T2_FK)
REFERENCES dbo.Table1
(T1_PK)
create
Table dbo.Table3
T3_PK        int
NOT NULL
Identity(1, 1)
CONSTRAINT [PK_T3_Constraint]
PRIMARY KEY (T3_PK),
T3_Field1    varchar
NULL,
T3_Field2    varchar
NULL,
T3_FK        int
NOT NULL
CONSTRAINT [FK_T3_T2]
FOREIGN KEY (T3_FK)
REFERENCES dbo.Table2
(T2_PK)
create
Table dbo.Table4
T4_PK        int
NOT NULL
Identity(1, 1)
CONSTRAINT [PK_T4_Constraint]
PRIMARY KEY (T4_PK),
T4_Field1    varchar
NULL,
T4_Field2    varchar
NULL,
T4_FK        int
NOT NULL
CONSTRAINT [FK_T4_T3]
FOREIGN KEY (T4_FK)
REFERENCES dbo.Table3
(T3_PK)
GO
--Create a basic view to select some data
CREATE
VIEW dbo.AliasView1
AS
select
t2.T2_FK as Table2_ForeignKey,
t1.T1_Field1 as Table1_FieldOne,
t2.T2_Field1 as Table2_FieldOne
FROM Table1 t1
Left outer
join Table2 t2 on t2.T2_FK
= t1.T1_PK;
GO
--Create another view that select basic data, and also selecting data from view 1
CREATE
VIEW dbo.AliasView2
AS
select
v1.Table1_FieldOne
as Table1_FieldOne,
v1.Table2_FieldOne
as Table2_FieldOne,
t3.T3_Field1 as Table3_FieldOne,
t3.T3_Field2
FROM Table3 t3
Left outer
join AliasView1 v1 on v1.Table2_ForeignKey
= t3.T3_PK;
GO
--My attempt to get the desired output, but no luck
SELECT
col.COLUMN_NAME as AliasColumnName, col.DATA_TYPE, col.CHARACTER_MAXIMUM_LENGTH
as max_length, colu.*
FROM
information_schema.COLUMNS col
left
outer join
(SELECT
VIEW_SCHEMA, VIEW_NAME, COLUMN_NAME,
min(TABLE_NAME)
as TABLE_NAME
FROM information_schema.VIEW_COLUMN_USAGE colu
WHERE VIEW_NAME =
'AliasView2'
Group by VIEW_SCHEMA, VIEW_NAME, COLUMN_NAME
) COLU ON colU.VIEW_NAME
= col.TABLE_NAME
and colu.COLUMN_NAME
= col.COLUMN_NAME
left
outer join
(select a.name
as TableName, c.name
as FieldName
from sys.foreign_key_columns fk
join sys.all_objects a
on a.object_id
= fk.parent_object_id
join sys.all_columns c
on c.object_id
= a.object_id
and c.column_id
= fk.parent_column_id
join sys.all_objects ar
on ar.object_id
= fk.referenced_object_id
join sys.all_columns cr
on cr.object_id
= ar.object_id
and cr.column_id
= fk.referenced_column_id
join sys.schemas scr
on scr.schema_id
= ar.schema_id
) fks on fks.TableName
= colu.TABLE_NAME
and fks.FieldName
= colu.COLUMN_NAME
WHERE COL.TABLE_NAME
= 'AliasView2'
order
by col.ORDINAL_POSITION
This is the results being returned: (That is not 100% what I am looking for)
AliasColumnName
DATA_TYPE
max_length
VIEW_SCHEMA
VIEW_NAME
COLUMN_NAME
TABLE_NAME
Table1_FieldOne
varchar
1
dbo
AliasView2
Table1_FieldOne
AliasView1
Table2_FieldOne
varchar
1
dbo
AliasView2
Table2_FieldOne
AliasView1
Table3_FieldOne
varchar
1
NULL
NULL
NULL
NULL
T3_Field2
varchar
1
dbo
AliasView2
T3_Field2
Table3
The desired results must be like the following:
AliasColumnName
DATA_TYPE
max_length
VIEW_SCHEMA
VIEW_NAME
COLUMN_NAME
TABLE_NAME
Table1_FieldOne
varchar
1
dbo
AliasView2
T1_Field1
Table1
Table2_FieldOne
varchar
1
dbo
AliasView2
T2_Field1
Table2
Table3_FieldOne
varchar
1
dbo
AliasView2
T3_Field1
Table3
T3_Field2
varchar
1
dbo
AliasView2
T3_Field2
Table3
NOTE: the COLUMN_NAME and TABLE_NAME must the REAL field of the TABLE it belongs to and not only ONE LEVEL Higher’s ALIAS View Name

Now we have a lot of views which select data from the tables, but I need to get the "Table Column Name" that is linked in the view.
If you are using SQL Server 2012/2014, then you can use
sys.dm_exec_describe_first_result_set (Transact-SQL) to gte the informations.
Olaf Helper
[ Blog] [ Xing] [ MVP]

Source table has an xml column that I need to query before splitting into 2 tables.

Hi
I have a source table with an xml column that I need to query on the nodes and then depending on what data is in the node then to split it into 2 different tables. I don't think I can use conditional split so I am looking at just a straight Execute SQL Task.
Has anyone got any ideas on any other way of doing this?
Thanks for the help in advance.
Dave

Hi TheDish,
Your goal arguably can me accomplished using pure T-SQL only encompassing XPath or XQuery.
E.g. the XQuery would retrieve the nodes as explained in http://sqljudo.wordpress.com/2013/12/02/xpath-for-the-sql-server-dba/
And then you may want to use the WHERE clause to extract - insert the nodes that go to one or another table.
I don't see why SSIS is necessary, but you can run the same queries in two Data Sources, or Execute SQL Tasks (that capture a resultset).
Arthur
MyBlog
Twitter

Oracle Dictionary: check what table belongs a column from a view

hi,
Is there any way to check in the oracle data dictionary what is the table that a column in a view belongs to?
Imagine this
create view xpto as
select table_a.col1, table_a.col2, table_b.column1
from table_a, table_b
I want to know that column1 in the view xpto belongs to the table_b.
Is this possible? How?
Best Regards,
Joao Oliveira

Hi Mat,
I know that I can check the table columns in that view or in user_Tab_columns. The problem is that some views have columns from different tables that have the same name. Understood?
So I really need to know what tables are part of the view query.
Thanks
Joao

Weblogic LLR tables RECORDSTR column size issues

Hi ,
We have configured the LLR (Last logging resource to emulate 2 Phase commit XA ) non XA datasouce in our weblogic 10.3.5 domain and I see that weblogic internally creates WL_LLR_<servername> tables in the schema confgured with that LLR datasouce per managed server.
But during some transactions commit we are facing an error stating that RECORDSTR column size (which is default 1000 Bytes) is too small for the COMMIT RECORD. It works fine when we increased the RECORDSTR column to 2000 bytes . But the issue is what is the criteria to increase the column size , our application have a large global transactions and will 2000 bytes will be enough to hold that.
So my question is what actually gets stored within the LLR tables for a transaction , transaction logs or COMMIT record ? , what does COMMIT RECORD actually mean?
Is there a way to specify the table and column length of LLR Tables during creation of LLR data sources itself.

Maybe this can shed some light on things: http://docs.oracle.com/cd/E21764_01/web.1111/e13737/transactions.htm#i1145819

How to disable right click option re-size on table column header?

Hi All,
Please let us know how to disable right click option re-size on table column header.
The issue is that when I right click on the column header, the column is selected and the context menu with options like Sort, Columns, Resize Columns, etc.. is popping. we want to disable column re-size option.
We are binding the table values programatically (not using Bc4J) and the Jdeveloper version is 11.1.2.2
Thanks in advance,
- Vignesh S.

Hi Gawish,
Thanks for the reply.
This will make the particular column frozen and only work for that particular column.
My use case is that to remove the resize columns option from the context menu or to disable the right click option.
Making column selection as none will disable the right click option but we need column selection for sorting.
Is there any other way to achieve this?
Thanks in advance,
-Vignesh S.

I need recommendations in purchasing an external hard drive, to back up everything in my computer. What brand and Size?

I have a mid 2010 iMac, version 10.8.2, memory 4GB, 1TB sata disk.
Apple sent me an e-mail last year stating that my imac maybe affected to a recall. They said I have to back up my computer before sending it in to them. So its a great excuse to purchase an external hard drive.
But I am clueless. I would like to know what brand and size of memory is recommened for my imac and my needs. I will be mostly be backing up homevideos and alot of pictures, music, and some documents.
If someone could also explain about time machine? I guess its a whole separte thing from your media. I also would like to keep it affordable. I just purchased an external hard drive from www.macmall.com Seagate Backup Plus 1TB USB 3.0 Portable Drive - Black (STBU1000100).
But I don't know if that was a really great choice because for it to work with time machine I have to reformat certain things, and Im not sure if I would like to have that hassle.
I also wanted to know whats the difference between usb 3.0 and firewire. I read some forums saying that if you are to transfer videos you need to use firewire, and some agreed disagreed. Is that true?
Please help!?
I definitely would like to know what brands of external hard drives are highly recommended and good at price. The one I purchased is it any good?
Thank you!
-wjrlelidepaula

http://eshop.macsales.com/shop/firewire/1394/USB/EliteAL/eSATA_FW800_FW400_USB
http://www.lacie.com/us/products/product.htm?id=10554
http://www.newertech.com/storage/

Getting error while creating form and report on webservice: ORA-20001: Unable to create form on table. ORA-02263: need to specify the datatype for this column.

i am using the following description to create a web service reference:
web reference :REST
Name :Yahoo Map
URL :http://local.yahooapis.com/MapsService/V1/mapImage
HTTP Method: GET
Basic Authentication: No
Add Parameter:
Name       Type
appid        String
location    String
Output Format: XML
XPath to Output Parameters : /Result
Output Parameter:
Name       Path       Type
Url          /text()      String
Then i tried to create form and report on webservice:
Web Service Reference Type: Yahoo Map
Operation: doREST
All the fields i keep as default
I tick the checkbox (url)in report Parameter
After clicking next whereever required i click create button
I get the following error
ORA-20001: Unable to create form on table. ORA-02263: need to specify the datatype for this column.
Please someone help to solve this as i need to fix it urgently.

i exported the application from apex.oracle.com and imported it to our environment
import went fine, but when I ran the IR page I got
ORA-20001: get_dbms_sql_cursor error ORA-00904: : invalid identifier
evidently the problem is a lack of public execute on DBMS_LOB, which is used in the generated IR source.
while waiting for the DBA to grant privs on DBMS_LOB, changing the dbms_lob.getlength call to length() fixes the IR.
however, i am not getting the download link on the associated form page... changed templates, that's not the issue -- we'll see if that's a dbms_lob issue as well

APEX:Getting error while creating form and report on webservice: ORA-20001: Unable to create form on table. ORA-02263: need to specify the datatype for this column.

I am using Apex 4.2.2.00.11
am using the following description to create a web service reference:
web reference :REST
Name :Yahoo Map
URL :http://local.yahooapis.com/MapsService/V1/mapImage
HTTP Method: GET
Basic Authentication: No
Add Parameter:
Name       Type
appid        String
location    String
Output Format: XML
XPath to Output Parameters : /Result
Output Parameter:
Name       Path       Type
Url          /text()      String
Then i tried to create form and report on webservice:
Web Service Reference Type: Yahoo Map
Operation: doREST
All the fields i keep as default
I tick the checkbox (url)in report Parameter
After clicking next whereever required i click create button
I get the following error
ORA-20001: Unable to create form on table. ORA-02263: need to specify the datatype for this column.
Please someone help to solve this as i need to fix it urgently.

336554,
Looks like there is a 127-column limit on the number of report columns supported when using that wizard. Do you have more than that?
57434

I need to make a spreadsheet based on info from specific table columns and headlines above those tab

Hi all,
A client of mine is requesting that I create a spreadsheet from a huge catalog that I am doing for him (this will also be done for all similar huge catalogs). The spreadsheet needs to pull information from some of the table columns, as well as the headline for each table. (This spreadsheet will then go out to some other people who will be putting this information into an online database/website for the client.) Let me be more specific and show you an example below.
Information from InDesign Catalog:
Tea Set
Part #
MFG #
Description
123-456
654321
Tea Cup
789-010
010987
Saucer
Here is how the Excel Spreadsheet needs to look:
Part #
MFG #
Product Name
Description
Page #
123-456
654321
Tea Set
Tea Cup
2
789-010
010987
Tea Set
Saucer
2
You have no idea how I have been pulling my hair out over this. Doing this manually will be a nightmare each time. My first thought was that I could assign specific cell styles to the columns (cells) and then create a TOC and then finangle with that - well, that didn't quite work, especially because of the headline I needed to pull into the spreadsheet as well. Is there ANY way I can get this information onto a spreadsheet? Knowing how wonderful InDesign is and all the tremendous capability it has, I am sure there IS a way and I just dont know how. Please PLEASE can someone point me in the right direction?
Thank you very much in advance for your assistance:) Any help would be greatly appreciated.
Christine

Sorry for the delay.
I am sure that this can somehow be done by making use of the styles. For example, when creating the TOC InDesign does a GREAT job of pulling in the styles (that are of course associated with the headlines and whatever needs to be included in the TOC) and associating them with the correct page numbers. I had experimented a little with creating specific paragraph styles and working these into the table and cell styles...then applying the Table Style to the table. I was halfway getting to where I wanted to be by creating a TOC that pulled in all of these styles and applied the corresponding page numbers (that I could somehow work into converting back into a table to then flow onto an Excel spreadsheet) but the wrench in the works was of course the Product Name. That said, I am VERY confident that this CAN be done somehow, by making use of the styles, using the same (or similar) methodology that is used to create the TOC. If only I were a pro at scripting...
Well, it's great that you're certain and confident, but I think your confidence is misplaced.
I had sort of hoped that running headers could be convinced to support this, but that doesn't seem to be true.
I think you need to postprocess the data somehow. If I was doing this for real I'd probably export the document to IDML and have a seperate program read the IDML XML, find the product names and page numbers, and insert them in tables. Though...that might mess up the page numbers. And it'd be a lot of work.
An much-faster-to-rapid-prototype solution is to have an indesign script go through the document and add in page numbers and the product names. Here's a rough prototype:
var d,s,story,table,product,page,c2,c3,c5,i,j.k;
d=app.activeDocument;
s=d.stories;
for (i=0; i<s.length; i++) {
    story = s[i];
    for (j=0; j<story.tables.length; j++) {
        table = story.tables[j];
        product = story.chara
cters[table.storyOffset.index-2].paragraphs[0].contents;
        page = table.parent.parentPage.name;
        // $.writeln("This is table "+j+". Product name: "+product+" page "+page);
        c2=table.columns[1];
        c3=table.rows[0].columns.add(LocationOptions.AFTER, c2);
        c3.contents=product;
        c5=table.rows[0].columns.add();
        c5.contents=page;
This doesn't handle the header rows in tables, but hopefully that's ok.
It may get page numbers wrong on multipage tables. Probably it would be better to just use
SpecialCharacters.AUTO_PAGE_NUMBER on the last full line (c5.contents=).
It should also probably not just look back 2 characters from the start of the table to find the paragraph, but it's probably good enough.

How to know columns and table name whose column size are modified

Hi guys
I want to know which all columns in the tables are modify
i.e. list of columns and table name whose column size are modified
Step1 :
CREATE TABLE employees
( employee_number number(5) ,
employee_name varchar2(50) ,
department_id number(10)
CREATE TABLE Supplier
( Supplier_number number(5) ,
Supplier_name varchar2(50) ,
CREATE TABLE customers
( customer_id number(10) not null,
customer_name varchar2(50),
address varchar2(50),
city varchar2(50),
state varchar2(25),
zip_code varchar2(10),
Step2 :
Alter table employees
MODIFY employee_number number(10)
ALTER TABLE supplier
MODIFY supplier_name varchar2(100)
step3
query to dispaly
columnname table name
employee_number employees
supplier_name supplier
How to know columns and table name whose column size are modified
could you please provide query
Thanks in Advance

09:35:50 SQL> desc dba_objects
Name                            Null?    Type
OWNER                                  VARCHAR2(30)
OBJECT_NAME                             VARCHAR2(128)
SUBOBJECT_NAME                         VARCHAR2(30)
OBJECT_ID                             NUMBER
DATA_OBJECT_ID                         NUMBER
OBJECT_TYPE                             VARCHAR2(19)
CREATED                             DATE
LAST_DDL_TIME                             DATE
TIMESTAMP                             VARCHAR2(19)
STATUS                              VARCHAR2(7)
TEMPORARY                             VARCHAR2(1)
GENERATED                             VARCHAR2(1)
SECONDARY                             VARCHAR2(1)
NAMESPACE                             NUMBER
EDITION_NAME                             VARCHAR2(30)LAST_DDL_TIME can be utilized

Change Font Size for Table Column

Hi,
How can we set different font sizes for table columns?
Thanks,
Uma.A

Hi
Set Design property of table column
Aslo check this link
http://help.sap.com/erp2005_ehp_03/helpdata/EN/66/18b44145143831e10000000a155106/frameset.htm
Thanks

Estimate index size on a table column before creating it

Hi
Is it possible to estimate the size of the index before actually creating it on a table column.
I tried the below query. It gives size of the index after creating it.
SELECT (SUM(bytes)/1048576)/1024 Gigs, segment_name
FROM user_extents
WHERE segment_name = 'IDX_NAME'
GROUP BY segment_name.
Can anyone through some light which system table will give this information.

You can get an approximation by estimating the number of rows to be indexes, the average column lengths of the columns in the index, and the overheads for an index entry - once you have some reasonable stats on the table.
I wrote a piece of code to demonstrate the method a few years ago - it's got some errors, but I've highlighted them in an update to the note: http://www.jlcomp.demon.co.uk/index_efficiency_2.html
Regards
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}{noformat} (lowercase, curly brackets, no spaces) so that the text appears in fixed format.
"Science is more than a body of knowledge; it is a way of thinking"
Carl Sagan

What is data being Skewed in a TABLE COLUMN in Oracle

Hi All,
I am going through numerous articles in Oracle to understand Cardinality,selectivity,histograms and Data being skewed to understand the characterstics of Oracle Optimizer.I have a hard time understanding the concept of data being skewed for a column in a table in Oracle.I need your help in understanding it.
Thanks,
Rav.

>
With out understanding the proper meaning of Skewness I assumed that a col data is skewed if the NDV of the col is high.
Your previoius statement raised this question now
Is there a corelation between number of distinct values in the column and the Skewness of a column?
>
See this Oracle white paper on Understanding Optimizer Statistics
http://www.oracle.com/technetwork/database/focus-areas/bi-datawarehousing/twp-optimizer-stats-concepts-110711-1354477.pdf
The section on Table and Column Statistics on page 2 has information that will clear this up
>
For example, if a table has 100 records, and the table access evaluates an equality predicate on a column that has 10 distinct values, then the Optimizer, assuming uniform data distribution, estimates the cardinality to be the number of rows in the table divided by the number of distinct values for the column or 100/10 = 10.
Histograms tell the Optimizer about the distribution of data within a column. By default (without a histogram), the Optimizer assumes a uniform distribution of rows across the distinct values in a column. As described above, the Optimizer calculates the cardinality for an equality predicate by dividing the total number of rows in the table by the number of distinct values in the column used in the equality predicate. If the data distribution in that column is not uniform (i.e., a data skew) then the cardinality estimate will be incorrect. In order to accurately reflect a non-uniform data distribution, a histogram is required on the column. The presence of a histogram changes the formula used by the Optimizer to estimate the cardinality, and allows it to generate a more accurate execution plan.
>
Did you notice the phrase 'if the data distribution in that column is not uniform (i.e. a data skew)'?
The NDV doesn't matter. It is the relative magnitude of each distinct value that determines the skew.
SB's example of gender, male or female, is not skewed if the number of 'males' is approximately equal to the number of 'females'.
But if the number of one of them, 'male' for example, is substantially larger than the number of the other then the data is skewed.
Assume a table has 1 million rows. If only one row is 'male' it makes sense to use an index to find it. But it wouldn't make sense to use that index if you were looking for 'female' since all rows but one are 'female'.
But because the NDV is two and there are 1 million rows then without a histogram Oracle would assume a uniform distribution and the query 'WHERE GENDER = 'male' would use a full table scan instead of the index.

What table column size is needed to accomodate Unicode characters

Similar Messages

Maybe you are looking for