Foreign key as part of primary key - bad idea

What are the pros and cons of using the primary key of the parent table as part of the primary key in the child tables versus just having it as a foreign key in the child tables?
Please share your experience on this. Thanks in advance.
Regards,
Richard

sb92075 wrote:
but if the relacion is 0-1 to 1-1 ; better design 2 tablesWhy?
Blank columns require no special accommodation.Like most things, there's usually a case (even if it's a small subset of all possible cases) where this makes sense.
http://www.oracle.com/technology/oramag/oracle/09-mar/o29asktom.html “Wide Load Storage”
Would be one such case.

Similar Messages

Use of end-date as part of Primary key in SAP Tables

All,
Any ideas on the rationale behind SAPs use of end-date as part of Primary key in general, specifically on condition Tables(A9xx for example) instead of Start-date? appreciate any help!!
Best Regards
TRP

Hmm, another example is CSKS, this logic was probably designed already in the 1980s...maybe it was seen beneficial to be able to select the currently valid record by using the primary key only (the first record where DATBI is larger than system date, not possible when DATAB is the key instead).
Maybe there are better explanations. Why do you need to know this, just curious?
Thomas

Making existing fields part of primary key

Hi,
There is a Z table(custom table), which has 3000 records. There are 12 fields in the table out of which 6 of them form the primary key. I need to make 2 more fields in the table part of primary key.
Is that possible? Will it affect the existing data in the table?
Please explain the possible ways of doing it.
Points will be given to all the helful answers.
Thanks- Nobin

Hi,
There can never be a possibility of Data issues. If a record in a table becomes unique by just the 6 fields , then the existing data will remain unique even if you add 2 more fields in the key.
Data issues can only come if you are reducing the key fields.
Take an example of a table with 2 fields and the data as show below
A1 B1
A1 B2
A2 B1
A2 B2
Now in this case if you reduce the key fields and make the first column only as the key then there can be issues as the there will be multiple rows with the same key.
In case you increase the key fields , then since the data was unique, the data will remain unique now also.
A1 B1 C1
A1 B2 C1
A2 B1 C1
A2 B2 C1
Regards,
Aj

How to khow how many foreign key are referenceing the primary key

sir i would like to know that how many foreign key are referencing the primary key of a table
is any command to know the table name
please suggest
thanks in advance..

In the query below, table_name is the name of the table whose primary key is referenced.
SELECT b.owner owner
,b.constraint_name foreign_key
,b.constraint_type constraint_type
,b.table_name referencing_table
,b.r_constraint_name primary_key
,c.table_name referenced_table
,c.column_name referenced_column
FROM all_constraints a
,all_constraints b
,all_cons_columns c
WHERE a.owner = b.owner
AND a.table_name = :table_name
AND a.constraint_type = 'P'
AND b.r_constraint_name = a.constraint_name
AND b.constraint_type = 'R'
AND c.constraint_name = a.constraint_name
AND c.owner = a.owner;
If the name of the primary key is known, the query below
select * from all_constraints where constraint_type = 'R' and r_constraint_name = :pk_name

Primary Key Materlized View and primary key in master tables

We are using Oracle 11g2 and I created a materlized view on several master tables. The master tables all have primary key, but I did not include all primary key columns in the defining query.
According to Oracle Manual e10592 Specify WITH PRIMARY KEY to create a primary key materialized view. This is the default ....
The master table must contain an enabled primary key constraint, and the defining query of the materialized view
must specify all of the primary key columns directly.My MV created OK and works. My question is what can be the impact on my MV? Could it hit the performance. It takes 40 minutes for the MV to be created and 5 minutes to refresh for single row insert in each mater table. I would like to see it refresh in seconds.

Retrieving rows from the mview is purely a function of the number of rows and indexing on the mview itself. Refreshing the mview wall time depends on the quality of the query used to refresh the mview. Whether you include PKs from the source table in the mview is of little, or no, consequence here. What matters is the execution path of the refresh query. Is there proper indexing on the underlying tables to support the joins. Did you use mlog$_s to enable fast refresh - are mlog$_s even usable for your mview. There are a number of possible causes to your problem and we do not have enough information to make an estimate of the problem.
BTW, if the mview could refresh in seconds w/o mlog$_s, you probably do not need an mview.

Unique key violation i.e. primary key violation in JSF page

Hi, I am using Jdeveloper 11.1.1.5.0 and working with adf fusion web application. I have created entity object and view object. In view object I have primary key and I have set it's type to DB sequence. THe problem is though the attribute is DB sequence. IN JSF page it show unique key violation i.e. primary key violation. Please help me

So, do you have a trigger in the database that updates the value of the primary key? Or do you assign the value in some other way?
You can check [url http://download.oracle.com/docs/cd/E16162_01/web.1112/e16182/toc.htm]the docs for information about the ways it can be done.

Setting CMP Primary Key equal to DB primary key

Is their a mechanism in Oracle JDeveloper(or other tools) for setting the primary key of auto-generated CMP beans to equal the database primary key?
If so, could someone kindly share their views on the benefits/drawbacks of this approach to CMP?
-Teodoro Sorgo

Brice,
It sounds like you have a page process on every page. Why not just wait to process all of the data on the last page, using your own pl/sql process? Say you have pages 1-3 with items
P1_ID
P1_NAME
P2_DEPT
P2_LOCATION
P3_OTHER
Only have a process on Page 3 have that does
insert into table (id, name, dept, location, other) values
(:P1_ID, :P1_NAME, :P2_DEPT ...
Anton

How to create a primary key or foreign key

in the table student, sid should be the primary key
in the table prerequisite cnum should be the foreign key and prereq the primary key
please help me with the syntax thanks.

As I said in your other post, it's all in the docs, go first to http://tahiti.oracle.com and try to find the answer yourself
http://download-uk.oracle.com/docs/cd/B10501_01/server.920/a96540/clauses3a.htm#1002630

Overwriting a record with part of same primary key

Hello Everyone,
Could anyone please give a suggestion about how to overwrite the record with part of same primary key?
I have 5 key fields.
A1
A2
A3
A4
A5
I have to overwrite the record if A5 changes with same A1 A2 A3 A4.
I cannot take off A5 as part of primary key. If I use modify or insert, its creating a new record with the new A5. But I am still having the old record in the table (with old A5), which I don't want it to be there.
Thank you,
Rushi

hi,
That's right.
u have to delete record from database table .....
then u can add a new record...
use delete <database_table> from table <internal_table>

Two foreign keys reference on primary key

There are two tables:
1) table CALENDAR with primary key: cal_id
2) table FACTS with some columns, two of them are dates: cal_id_start_process, cal_id_stop_process
we want on physical layer create two foreign keys for cal_id_start_process and cal_id_stop_process columns reference on CALENDAR.cal_id
When we create foreign key for cal_id_start_process: we choose this column and choose CALENDAR table and its primary key. But when we want to create foreign key for cal_id_stop_process - we cannot coose table CALENDAR again.
Is this rigth, that only one column from table can reference on specified primary key in CALENDAR table?
How to create two foreign key on one table primary key?

More complex. In your example there'll be such SQL:
select c.cal_id, f.cal_id_start, f.cal_id_stop
from CALENDAR c, FACTS f
where c.cal_id = f.cal_id_start and c.cal_id = f.cal_id_stop
(It meens FACTS.cal_id_start = FACTS.cal_id_stop )
In my case, I want for such join:
select c1.cal_id, c2.cal_id, f.cal_id_start, f.cal_id_stop
from CALENDAR c1, CALENDAR c2, FACTS f
where c1.cal_id = f.cal_id_start and c2.cal_id = f.cal_id_stop
Edited by: annylut on Dec 22, 2011 3:33 PM

Primary keys and foreign keys

Hi,
is there a script available for changing the value of a primary key which automaticly
detect all foreigs keys refering to this primary key and change them as well?
William Oorschot

I don't have the script now. but you can make a procedure of ur own.
using a join on user_constraints, and user_cons_columns you can find out all the columns in all the tables which are referencing the particular primary key. once you have the table and the referencing columns, you can change the value of those columns with the new value.
you'll have to call this procedure from a before row level trigger.
HTH
Naveen

Buffer Busy Waits on Primary Keys - RKI

11.2.03 - HPUX
Hi All.
We're seeing significant Buffer Busy Waits (around 15% of non idle waits during bad times) and most of the contention is on the following;
SYS.SEQ$, and two primary keys..
The primary key one seems easier to fix in the first instance, with most of the expert thinking seeming to be either Reverse Key Indexes, or hash partitions for index. Thinking this through, wouldn't moving from normal key order to reverse key order cause a lot more I/O - if the assumption is that the rows being queried are likely to be the newer rows being inserted?? If we scatter them evenly among the index segment - won't that have a deleterious effect on I/O for that segment? In effect we're swapping Buffer Busy Waits for much worse (?) I/O?

>
We're seeing significant Buffer Busy Waits (around 15% of non idle waits during bad times) and most of the contention is on the following;
SYS.SEQ$, and two primary keys..
The primary key one seems easier to fix in the first instance, with most of the expert thinking seeming to be either Reverse Key Indexes, or hash partitions for index. Thinking this through, wouldn't moving from normal key order to reverse key order cause a lot more I/O - if the assumption is that the rows being queried are likely to be the newer rows being inserted?? If we scatter them evenly among the index segment - won't that have a deleterious effect on I/O for that segment? In effect we're swapping Buffer Busy Waits for much worse (?) I/O?
>
If you have heavy INSERTs, especially in RAC, you need to solve your current problem. But you need to make sure you know what that problem is.
You have likely identified it but you should take a quick look at your sequence and the cache value being used. Contention on SEQ$ could mean that you are using NOCACHE or a low value for CACHE.
If your primary key accesses typically access just one row it won't really matter if those rows are 'scattered'. It sounds like you are thinking that queries of 'newer' rows will make it likely that those queries will need the same blocks because the rows will be sharing a block.
But why would users be accessing data using a non-meaningful sequence number? I wouldn't expect users to even know what the primary key value was to be able to query it. I would expect them to be querying based on a business value, company_code or date.
Can you explain why you are worried about that?
You might be interested in this two part article about reverse key indexes by Richard Foote
http://richardfoote.wordpress.com/2008/01/14/introduction-to-reverse-key-indexes-part-i/
http://richardfoote.wordpress.com/2008/01/16/introduction-to-reverse-key-indexes-part-ii-another-myth-bites-the-dust/
>
Reverse Key Indexes are designed to resolve a specific issue, that being index block contention. Many indexes in busy database environments with lots of concurrent inserts (and in some scenarios updates and deletes as well) can suffer from index block contention (as highlighted by high levels of “buffer busy waits” and “read by other session” wait events for the index segments). Monotonically increasing indexes, such as Primary Keys generated by a sequence, are especially prone to contention as all inserts need to access the maximum “right-most” leaf block. This is of particular concern in RAC environments, where this “hot” index block needs to be accessed by all the instances and is being bounced around the various SGAs causing expensive block transfers between instances.
A solution is make the index a Reverse Key Index.

What is the best practice for creating primary key on fact table?

what is the best practice for primary key on fact table?
1. Using composite key
2. Create a surrogate key
3. No primary key
In document, i can only find "From a modeling standpoint, the primary key of the fact table is usually a composite key that is made up of all of its foreign keys."
http://download.oracle.com/docs/cd/E11882_01/server.112/e16579/logical.htm#i1006423
I also found a relevant thread states that primary key on fact table is necessary.
Primary Key on Fact Table.
But, if no business requires the uniqueness of the records and there is no materilized view, do we still need primary key? is there any other bad affect if there is no primary key on fact table? and any benifits from not creating primary key?

Well, natural combination of dimensions connected to the fact would be a natural primary key and it would be composite.
Having an artificial PK might simplify things a bit.
Having no PK leads to a major mess. Fact should represent a business transaction, or some general event. If you're loading data you want to be able to identify the records that are processed. Also without PK if you forget to make an unique key the access to this fact table will be slow. Plus, having no PK will mean that if you want to used different tools, like Data Modeller in Jbuilder or OWB insert / update functionality it won't function, since there's no PK. Defining a PK for every table is a good practice. Not defining PK is asking for a load of problems, from performance to functionality and data quality.
Edited by: Cortanamo on 16.12.2010 07:12

Help.......... required :Database table with no primary key

Hi All,
I have a table sponsor_logos whose structure is as follows:
sl__sponsorid number,
sl__clubid number,
sl_hallno varchar2(7) not null,
sl_logofilename varchar2(100) not null,
sl_date_modified date not null
Where in the sl__sponsorid and sl__clubid are foreign keys and I do not have any primary key for the table.
I need to insert into and also delete records from the table.
I have gone through various topics in the java forum for EJB's to find out what design strategy should I use?
Some say that if a table does not have primary keys, and if the table is represented by a CMP entity bean all the fields of the table are considered as a part of primary key.In that case what would my code look like and what would my deployment descriptor contain?????
According to the specification, the CMP or a BMP represents a unique record in a database.In that case should I be using a Session bean to handle insertions and deletions for the table?
Since this is my first project with EJB's I am really confused as to what design should I be using for this situation. Since I have a time constraint I can't take a long time to decide on what strategy I should use.
Suggestions regarding this are welcome .

Hi Maris,
I have my PK class ready with all the fields in it.
Deployment descriptor ready with <prim-key-class>package.SponsorLogoPK</prim-key-class>
. According to descriptor specification there is no need to have an entry for
<primkey-field>bookingSlNo</primkey-field>
in case of custom primary key .
So I have omitted <primkey-field>bookingSlNo</primkey-field>
I have used the verifier tool and verified the ejb , no errors with regards to this ejb, but , getting errors while deploying:
An error has occurred.
Deployment Error -- Error while running ejbc -- Fatal Error from EJB Compiler -- JDO74025: JDOCodeGenerator: Caught an Exception validating CMP bean 'SponsorLogoBean' in application 'resnet_ejb' module 'E:\Sun\AppServer\domains\domain1\applications\j2ee-modules\resnet': JDO72335: If the table SPONSOR_LOGOS for the bean corresponding to the generated class com.resnet.squash.api.entity.SponsorLogoBean483209734_JDOState is mapped as the primary table, it must have a primary key. Choose a different primary table or verify that the contents of the schema file are correct.
Now what could the problem be . Do I need to do any other setting in the descriptor. I have the cmp-mapping also intact.

Master Detail Forms with 2 composite primary keys - Is there a workaround?

Hello All,
I have been searching for a workaround to the maximum 2 part primary key restriction on the multi-row updates, and master-detail forms, and am hoping that someone can help me. I am using HTMLDB v2.0.0.00.49 with IE 6 against a 9.2 DB.
I successfully implemented the workaround of Fred Stoopendaal's (see Updata PK on HTML DB ) and it works fine for single page multi-record updateable forms, but alas I haven't been able to extend it to master detail forms (I think it is something to do with Oracle not allowing the "returning" clause on views).
Here is what I tried:
two tables, one with a 2 part composite primary key, which is the master table, and a detail table with 3 part composite primary key -
--------- BEGIN SQL ---------
create table master_table
( master_col1 number
, master_col2 number
, master_col3 varchar2(30)
, constraint master_pk primary key (master_col1,master_col2));
create table detail_table
(detail_col1 number
,detail_col2 number
,detail_col3 number
,detail_col4 varchar2(30)
, constraint detail_pk primary key(detail_col1,detail_col2,detail_col3)
, constraint master_detail_fk foreign key (detail_col1,detail_col2) references master_table(master_col1,master_col2));
create or replace view v_master_table as
select rowid mata_rowid,mata.*
from master_table mata;
create or replace view v_detail_table as
select rowid deta_rowid,
(select rowid from master_table mata where mata.master_col1 = deta.detail_col1 and mata.master_col2 = deta.detail_col2) deta_mata_rowid
, deta.*
from detail_table deta;
create or replace trigger mata_ins_upd_trg
instead of insert or update on v_master_table
referencing new as new old as old
for each row
begin
if inserting then
insert into master_table (master_col1, master_col2, master_col3)
values (:new.master_col1, :new.master_col2, :new.master_col3);
end if;
if updating then
update master_table
set master_col1 = :new.master_col1,
master_col2 = :new.master_col2,
master_col3 = :new.master_col3
where rowid = :old.mata_rowid;
end if;
end;
create or replace trigger deta_ins_upd_trg
instead of insert or update on v_detail_table
referencing new as new old as old
for each row
begin
if inserting then
insert into detail_table ( detail_col1, detail_col2, detail_col3, detail_col4)
values (:new.detail_col1, :new.detail_col2, :new.detail_col3, :new.detail_col4);
end if;
if updating then
update detail_table
set detail_col1 = :new.detail_col1,
detail_col2 = :new.detail_col2,
detail_col3 = :new.detail_col3,
detail_col4 = :new.detail_col4
where rowid = :old.deta_rowid;
end if;
end;
--------- END SQL ---------
Then I created a master-detail form in Apex on the two views, using the mata_rowid and deta_rowid as primary keys, and mata_rowid=deta_mata_rowid as the link. I realise that using a function to fetch the master rowid within the detail view query is costly, but it was my intention to modify the record fetch queries to use the real FK columns once things were up and running.
It seems to generate the pages ok, and I can insert/update master table records, but as soon as I modify records in the detail table things go a bit haywire. I can't find any documentation on how the inbuilt MRU/MRD logic works, so can't figure out the issue.
Can anyone out there tell me what the problem is with the logic above, or if they have come up with a neat solution to this annoying limitation. I know that many will say that I should modify the data model to use surrogate primary keys, but many of the uses for HTMLDB are new interfaces for old schemas, so a workaround that doesn't involve wholesale data model changes would be preferable.
Thanks in advance,
Mike Cretan

Hi, this is likely not the most elegant way...but perhaps the simplest -- and I didn't have much time to play.
I used Wizard to create two separate Master Detail forms, each with a separate detail table. Thus I ended up with four pages:
Page "A" - "Selector" page for Master (Report), with Edit link driving to Detail-1
Page "B" - Editable Master/Detail-1 page (HTML / Report)
Page "C" - "Selector" page for Master (Report), with Edit link driving to Detail-2
Page "D" - Editable Master/Detail-2 page (HTML / Report)
Then I selected the primary key column TWICE on the Report on Page A. Modified the second instance of this column to navigate to Page D (passing primary key) exactly the way the original instance of this column navigates to Page B. Then I deleted Page C.
Since you can have only one Tabular Entry form per page, this seemed the best way to drive two separate detail tables from a common interface.

Foreign key as part of primary key - bad idea

Similar Messages

Maybe you are looking for