Functions of outer join left as simple join
I created two views SCO_REQGROSSLINES_V and SCO_REQLINESCOMPLETE on Oracle applications tables base. Our company applies the following identity rules: the gross number of rows is always greater than or equal number of demand lines. (Full, I mean delivered.)Each view has a key named DATESUBINV_KEY built a date of the operation and the code subinventory using this logic:
TO_CHAR (TRUNC (transaction_date), 'YYYYMMDD') | subinventory_code
On 03-SEP-2010 a user has created a demand with a line item for the material of the HEAT subinventory who did not. This row is returned in the result set when I ask about SCO_REQGROSSLINES_V but does NOT return when I ask about SCO_REQLINESCOMPLETE. So far so good.
When I run either of the left outer joins following, I don't get the line for the line for HEAT subinventory 03-SEP-2010:
Select rgl.trxn_date
rgl.subinventory
rgl.line_count gross_line_count
rgl.ext_prc gross_value
rlc.line_count lines_filled
rlc.ext_prc lines_filled_value
of SCO_REQGROSSLINES_V vö outer join left sco_reqlinescomplete_v rlc
ON rgl.datesubinv_key = rlc.datesubinv_key
where vo. TRXN_DATE BETWEEN to_date('29-AUG-2010') AND to_date('04-SEP-2010')
and rlc.trxn_date BETWEEN to_date('29-AUG-2010') AND to_date('04-SEP-2010')
order of rgl.trxn_date, rgl.subinventory
Select rgl.trxn_date, rgl.subinventory
rgl.line_count gross_line_count
rgl.ext_prc gross_value
rlc.line_count lines_filled
rlc.ext_prc lines_filled_value
of SCO_REQGROSSLINES_V vo
sco_reqlinescomplete_v rlc
where vo. TRXN_DATE BETWEEN to_date('29-AUG-2010') AND to_date('04-SEP-2010')
and rlc.trxn_date BETWEEN to_date('29-AUG-2010') AND to_date('04-SEP-2010')
AND rgl.datesubinv_key = rlc.datesubinv_key (+)
order of rgl.trxn_date, rgl.subinventory
Question: Why the left outer joins is would work as a simple join on DATESUBINV_KEY? Do the keys function differently from the keys defined in the data dictionary views? If they do how these keys work as data dictionary keys?
Hello
The WHERE clause is applied after the outer join is complete. In your example, Vo will be in the game even if they have no match in rlc. In this case, all columns of rlc will be NULL.
Okay, so you have these lines containing data from vo but NULL values, where the rlc data was supposed to go. Now it's time to start the WHERE clause. You then apply that conditions in the WHERE clause:
AND rlc.trxn_date BETWEEN ...
NULL is between what whether, if the line the prodeuced the outer join is rejected.
All the columns in involivn conditions of rlc probably needs to be part of the join condition, not the WHERE clause.
For example, in you first ask the FROM and WHERE clauses can be:
FROM sco_reqgrosslines_v rgl
LEFT OUTER JOIN sco_reqlinescomplete_v rlc ON rgl.datesubinv_key = rlc.datesubinv_key
AND rlc.trxn_date BETWEEN TO_DATE ( '29-AUG-2010'
, 'DD-MON-YYYY'
)
AND TO_DATE ( '04-SEP-2010'
, 'DD-MON-YYYY'
)
WHERE rgl.trxn_date BETWEEN TO_DATE ( '29-AUG-2010'
, 'DD-MON-YYYY'
)
AND TO_DATE ( '04-SEP-2010'
, 'DD-MON-YYYY'
)
Always use (at least) two arguments when calling TO_DATE.
Conditions that involve only VO can be in the WHERE clause, but if all the rlc conditions in the WHERE clause, then the effect will be the same as a join internally.
It is no difference between tables and views (or subqueries) in this regard: the same thing would happen, whether it is vo and rlc really tables or not.
Tags: Database
Similar Questions
-
Hi Experts,
I have a requirement that says - see the chart for the past 10 days, regardless the presence data table in fact.
Lets consider an example - Time_dim product, are my dimension tables, Purchase_Order is my fact table.I did it for external Purchase_Order in left RPD with TIME_DIM and inner join with the PRODUCT table. and execution of query of exit-
Select T.Date, P.item, count (distinct PO.order_no)
TIME_DIM t, PRODUCT P, PURCHASE_ORDER PO
where T.date_key = PO.date_key
and P.item = in. agenda
and P.item = 'laptop ';The query generated by OBIEE left outer join, but when the condition P.item = "Notebook" included in the query, and if there are no orders for this product in one of the date, that date will not come in the result set.
the query to be generated by the OBIEE is-
Select T.Date, PO.item, count (distinct PO.order_no)
TIME_DIM t,.
(SELECT P.ITEM, IN. ORDER_NO
PRODUCT P, PO PURCHASE_ORDER
WHERE P.item = in. agenda
and P.item = 'Laptop') IN.
WHERE T.date_key = PO.date_key (+);How to design the RPD to achieve this. All pray to advise on this. Thanks in advance.
Thank you
ChantalHello
You are on 11.1.1.7?
I would say that your condition can be made without using external and maintenance of product and the standard between the FACT dimension, time inner join join.
If you enable your property analysis OBIEE "Include Null values" will automatically return all the elements of time and product matching your filter (so you'll need to add a filter on 'Date' to limit it to the last 10 days or you will have a unique day of your time dimension).
If you filter then on "Laptop", even if there is not a single value in order for "Laptop" in the last 10 days, he will be there on the screen.
Easy, clean and you keep your inner join between the facts and Dimensions.
Take a look at this example, I just did on SampleApp 406:
Selection of 12 months (year 2010) and a customer (id = 89) and income. The model has only an inner join. I activate the option "Include Null values" and here is the result.
A line with cells only empty because there is not a single revenue for customer 89 in 2010. This is exactly your condition.
Honestly, do not touch your model using the outer join, you will have more side effects than benefits. Every single scan will do the outer join and you'll have a lot of data 'empty' return of the DB (more large data set containing just the null values) and probably you need the outer join in 15 to 25% of your analysis.
Keep things simple, it will be faster and easier to maintain.
-
Modeling of the left outer join
Hello world
I'm tender hand to you guys for a modeling help
I have a FACT, the customers, the Dim_Date and CUST_ADDRESS of tables to model
Fact and the client are joined through CUST_ID
FACT and DATE are joined through DATE_ID
CUST_ADDRESS must be attached to the top of the model through CUST_ID, DATE_ID and this join must be Left outer because sometimes the address does not exist or is not current, which means DATE_ID could be different between Dim_Date and CUST_ADDRESS
If it were to join internal, model would have been easy, because of the outside left that I am unable to model, it's pretty good.
Application under
Select D.DATE, C.CUST_NAME, CA. ADDRESS, F.AMOUNT
Of
F FACT
JOIN THE
CUSTOMER C
ON C.CUST_ID = F.CUST_ID
JOIN THE
DIM_DATE D
ON F.DATE_ID = D.DATE_ID
LEFT OUTER JOIN
CUST_ADDRESS CA
ON C.CUST_ID = CA. CUST_ID AND C.DATE_ID = D.DATE_IDThanks in advance
When I add the CUSTOMER and in FACT LTS CUST_ADDRESS
Stop it!
Don't add CUSTOMER and CUST_ADDRESS in the FACT of LTS. Why would add you to the LTS DO?
You design a management model: CUSTOMER is a dimension and it has its own logical table this logic table join with a logical join in the activity diagram. Ditto for CUST_ADDRESS.
So the change, I missed earlier is CUST_ADDRESS contains no Cust_ID (ACTUALLY existing), but contains a Cust_NO, and the table to translate Cust_NO in Cust_ID is CUSTOMER?
No problem...
Let's start with a new alias of CUSTOMER (to keep more simple to understand at the moment), call as you want, but this new alias will be the link between the FACT and CUST_ADDRESS.
In LTS of the dimension 'Address', you have CUST_ADDRESS initially, add an inner join on the new alias that you created in the LTS of the CUSTOMER. So now your 'Address' logical dimension contains the Cust_NO and Cust_ID and this will make the join to FACT.
Between CUST_ADDRESS and the CLIENT, you can keep an inner join, because the target is not not for get the address of the customer, but is having the Cust_ID in the address line.
Give it a try at that.
But do not add these tables in the LTS, they are logical dimensions.
-
Help for a LEFT OUTER JOIN query
Hello, all,.
I'm having some trouble setting up an Oracle 11 g Server SQL query, and I could use some help.
Let's say tableA is blogs; tableC is comments for blog entries; tableB is the associative array:
tableA blogID blogTitle blogBody dateEntered 1 This is a test More text... 2016-05-20 11:11:11 2 More testing Still more! 2016-05-19 10:10:10 3 Third charm!! Blah, blah. 2016-05-18 09:09:09
tableC commID userID userText dateEntered 10 Bravo I like it! 2016-05-20 11:21:31 11 Charlie I don't! 2016-05-20 11:31:51 12 Alpha Do it again! 2016-05-19 10:20:30 13 Bravo Still more? 2016-05-19 10:30:50 14 Charlie So, what? 2016-05-19 10:35:45 15 Bravo Blah, what? 2016-05-18 09:10:11 16 Alpha Magic number! 2016-05-18 09:11:13
tableB blogID commID 1 10 1 11 1 12 2 13 2 14 3 15 3 16
I'm trying to get blogID, blogTitle, blogBody and the number of comments for each blog entry. But, since I'm on to_char() for date and COUNT (commID) for the total number of comments, I am not "a group by expression.
Here is an example of pseudo-SQL of what I'm trying.
SELECT a.blogID, a.blogTitle, a.blogBody, to_char(a.dateEntered,'YYYY-MM-DD HH24:MI:SS') as dateEntered, COUNT(c.commID) as total FROM tableA a LEFT OUTER JOIN tableB b ON b.blog_ID = a.blog_ID LEFT OUTER JOIN tableC c ON c.commID = b.commID WHERE a.blogID = '1' GROUP BY blogID, blogTitle, blogBody ORDER BY to_date(dateEntered,'MM-DD-YYYY HH24:MI:SS') desc
I'm sure it's something simple, but I just DO NOT see it. Can you help me?
V/r,
^_^
Try:
GROUP BY a.blogID, a.blogTitle, a.blogBody, to_char(a.dateEntered,'YYYY-MM-DD HH24:MI:SS')
See you soon
Eddie
-
Why left outer join with a table gives me more lines?
Hi gurus,
I can see "view_a" and a table 'table_a '.
view_a a county of 100 lines. Now, when I left outer join that discovers with a 'table_a', I expect all 100 lines.
However, I'm more than 100 lines. Is it still possible?
Also even to analyze these situations, how can I move forward?
Because it is very high volumn of sight and takes longer to run.
Select count (*) view_a, view_b
where view_a.col1 = view_b.col1 (+)
and view_a.col2 = view_b.col2 (+);
Thank you
I can see "view_a" and a table 'table_a '.
view_a a county of 100 lines. Now, when I left outer join that discovers with a 'table_a', I expect all 100 lines.
However, I'm more than 100 lines. Is it still possible?
Also even to analyze these situations, how can I move forward?
Because it is very high volumn of sight and takes longer to run.
Select count (*) view_a, view_b
where view_a.col1 = view_b.col1 (+)
and view_a.col2 = view_b.col2 (+);
Which is not necessarily related to the use of an outer join.
Just join of two tables in general will give you more rows of one table has.
Scott DEPT table contains ONE row for deptno = 10
The EMP table has THREE rows of deptno = 10
The number of rows you plan if you join two tables using an equi-join?
Three - what is MORE lines the DEPT table has for deptno = 10
Select * from Department where deptno = 10
DEPTNO, DNAME, LOC
10, ACCOUNTING, NEW YORKSelect * from emp where deptno = 10
MGR, EMPLOYMENT ENAME, EMPNO, HIREDATE, SAL, COMM, DEPTNO
7782, CLARK, MANAGER, 7839, 6/9/1981,2450, 10
7839, KING, PRESIDENT, 17 NOVEMBER 00, 10
7934, MILLER, CLERK, 7782, 23 JANUARY 00: 10Select dept.*, emp.*
Department, emp
where dept.deptno = 10
and dept.deptno = emp.deptnoDEPTNO, DNAME, LOC, EMPNO, ENAME, JOB, MGR, HIREDATE, SAL, COMM, DEPTNO_1
10, ACCOUNTING, NEW YORK, 7782, CLARK, MANAGER, 7839, 6/9/1981,2450, 10
10, ACCOUNTING, NEW YORK, 7839, KING, PRESIDENT, 17 NOVEMBER 00, 10
10, ACCOUNTING, NEW YORK, 7934, MILLER, CLERK, 7782, 23 JANUARY 00: 10So if these are the lines ONLY in the table EMP and DEPT the query would give you THREE lines despite the DEPT table only ONE line.
No do you expect? You get ALL the child rows that belong to the parent company. Otherwise, how could it possibly work?
The OUTER join includes lines where the parent row exists but there is NO child line as others have shown.
Outer joins
Outer join extends the result of a simple join. Outer join returns all rows that satisfy the join condition and also returns some or all rows in a table for which no line of the other meet the join condition.
Get more lines to exist in one of the paintings is a basic necessity. It usually has NOTHING to with the question of whether you have an outside to join or not.
See the section on the JOINTS in the Oracle documentation
http://docs.Oracle.com/CD/B28359_01/server.111/b28286/queries006.htm
-
Outer joins to the left... Please help!
Hmm. Ive had trouble with this one for a while and have read and re-read docs etc.
IM under 11.2 XE and its Apex 4.2 application...
Its a simple query with outer joins... but I can't simply to do it right after several hours of trying! ID be very grateful if anyone can lend a hand...
For the query below. I am looking for the following:
Complete list of entities, a sum of the values if they exist (and obviously empty if they don't), filtered by a table of SUM choice only the records that match a value in the lookup table.
Select
e.ENTITY as ENTITY,
Sum (p.amount) as Forecast_Income,
Sum (BR.bri_credit) as Actual_Income,
Sum (BR.bri_debit) as Actual_Expenses
Of
e entities
LEFT OUTER JOIN payments p
WE (e.entity_id = p.entity_id)
LEFT OUTER JOIN pmt_stat_lookup ps
WE (p.status_id = ps.status_id
and ps.forecast = 'Y')
LEFT OUTER JOIN bri_recon br
WE (e.entity_id = br.entity_id)
E.entity group
My problem with the join where I join p.status_id for ps.status_id and ps.forecast = 'Y '... If I run the above statement, it seems does not account for this filter criteria and summarizes all records of PAYMENTS, regardless of... If I move outside the outer join clause, it only brings back the records of PAYMENTS where there is a join. Ive tried to join in a different order etc... but in vain...
If someone could help, Id be very grateful.
Kind regards
Richard
Hello Richard
Try this query:
SELECT e.entity as ENTITY
sum (PO.amount) as Forecast_Income
sum (br.bri_credit) as Actual_Income
sum (br.bri_debit) as Actual_Expenses
Of
E ENTITIES
LEFT OUTER
JOIN (SELECT P.entity_id )
P.amount
PAYMENTS P
JOIN PMT_STAT_LOOKUP PS
ON p.status_id = ps.status_id
AND ps.forecast = 'Y '.
) IN.
WE e.entity_id = in. entity_id
LEFT OUTER
JOIN BR BRI_RECON
ON e.entity_id = br.entity_id
E.entity GROUP
;
I hope it helps.
Best regards, David
Post edited by: David Berger
-
Show all dates between date range (time Dimension is left outer join)
All,
I did some research on this issue, but in all positions on date variables, date prompts and date filtering I have not seen one exactly to my question (perhaps that they are and I don't have my head around it properly yet).
My requirement of report is to allow a user to select a start date and an end of day. The report is expected to show the activity of these two days - AND display 0/null on days where there is no activity. This second part is where I am getting hung up.
The paintings in question are:
Timedim
EventFact
CustomerDim
My MDB is configured as follows:
Left outer join of Timedim EventFact
Inner join CustomerDim EventFact
If I run a report by selecting the DAYS of Timedim and an EventFact measure1 with range day 01/01/2010-31/12/2010... A record for each day and it looks perfect because of the left outer join between Timedim and CustomerDim.
But... If I add a CustomerDim field, Select TimeDim.DAY, CustomerDim.CUSTNAME, EventFact.MEASURE1, OBIEE returns only records for the days that have record EventFact.
This is due to the fact that the Timedim is always external joined in EventFact, but adding in fact CustomerDim OBIEE set up an inner join between tables that will only return data where there are data EventFact.
There is a way around it in this simple case, and that is to define the relationship between CustomerDim and EventFact as an outer join as well. This will give the desired effect (but an outer join between the two tables is not the real relationship) and I have add an extra dimension and add additional sources of logic to a single dimension in MDB it becomes complicated and messy.
Also, ive ruined with the definition of the conduct in the relationship table, etc... but he gave not the desired effect.
Has anyone ever met the need for force display all dates within a range specified with a fact table that does not have an entry for each date?
Thanks in advance.
K
Published by: user_K on April 27, 2010 11:32Hi there, the easiest way is to the LTS himself. Double-click your LTS, go to the tab with the column mappings. Make sure you have checked "show no mapped" column.
You should see your new dummy column in the list, in the central part of mapping (IE not the right) just enter 0, or launch the expression editor and enter 0 in there.
simple! -
Problem format (LEFT OUTER JOIN)?
THE addresses of Mutiple of return as a single record + current addresses
>
Hi all
I have to back student addresses home and dormitory under a single registration.
He had to go something like this:
In this format the desired:LAST_NAME FIRST_NAME ADDY_TYPE ADDRESS ZIP Smith John HOME 123 Awesome St 10003 Smith John DORM Oak Quad 10013
You also need to get their last addresses by date.LAST_NAME FIRST_NAME ADDY_TYPE ADDRESS ZIP ADDY_TYPE ADDRESS ZIP Smith John HOME 123 Awesome St 10003 DORM Oak Quad 10013
The database hold records of all students have places moving from dorm to dorm
and some permanent residence ("HOME") has changed as well.
To return only a DORM address and only a HOME address
for each student.
I'm looking at possibly a function BOX to put on a line/record and RANK BY() or DENSE_RANK to determine the last addresses.
Hope I'm making some sense and as always very grateful for any help. Thank you.
>
The correct code provided by Frank Kulash here:
I need to add a 'NATION' field, located on another table for the addresses of welcome for foreign students.WITH got_rnum AS ( SELECT last_name , first_name , addy_type , address , zip , ROW_NUMBER () OVER ( PARTITION BY last_name , first_name , addy_type ORDER BY addy_date DESC NULLS LAST ) AS rnum FROM table_x -- WHERE ... -- Any filtering goes here ) SELECT last_name , first_name , MIN (CASE WHEN addy_type = 'HOME' THEN address END) AS home_address , MIN (CASE WHEN addy_type = 'HOME' THEN zip END) AS home_zip , MIN (CASE WHEN addy_type = 'DORM' THEN address END) AS dorm_address , MIN (CASE WHEN addy_type = 'DORM' THEN zip END) AS dorm_zip FROM got_rnum WHERE rnum = 1 GROUP BY last_name , first_name ;
I made a LEFT OUTER JOIN with the table of the NATION and the release came out like this:
My desired output would be like this:LAST_NAME FIRST_NAME ADDY_TYPE ADDRESS ZIP NATION ADDY_TYPE ADDRESS ZIP Smith John HOME Rue Henry M1V 4F4 CANADA null null null Smith John null null null null DORM Oak Quad 10013
Maybe it's something I'm not right. What is the way I'm joining tables?LAST_NAME FIRST_NAME ADDY_TYPE ADDRESS ZIP NATION ADDY_TYPE ADDRESS ZIP Smith John HOME Rue Henry M1V 4F4 CANADA DORM Oak Quad 10013
As always very grateful for your contributions. Thank you.Give a glance to your group by, you group on the nation, but it has two values (NULL and CANADA). Try this way:
WITH got_rnum AS ( SELECT STUINFO_id , STUINFO_last_name , STUINFO_first_name , ADDRESSLIST_atyp_code , ADDRESSLIST_street_line1 , ADDRESSLIST_street_line2 , ADDRESSLIST_city , ADDRESSLIST_stat_code , ADDRESSLIST_zip , ADDRESSLIST_natn_code , NATION_nation , ROW_NUMBER () OVER ( PARTITION BY STUINFO_last_name , STUINFO_first_name , ADDRESSLIST_atyp_code ORDER BY ADDRESSLIST_from_date DESC NULLS LAST ) AS rnum FROM STUINFO JOIN CLASSROSTER ON STUINFO_pidm = CLASSROSTER_pidm JOIN ADDRESSLIST ON ADDRESSLIST_pidm = STUINFO_pidm LEFT OUTER JOIN NATION ON ADDRESSLIST_NATN_CODE =NATION_CODE -- The WHERE part determines if the student is currently enrolled in a class -- ADDRESSLIST_to_date is the last date the student will be living in that residence WHERE ADDRESSLIST_atyp_code IN ('PR', 'CA') and STUINFO_change_ind IS NULL and STUINFO_last_name !='Registrar' and CLASSROSTER_term_code='200909' and CLASSROSTER_PTRM_CODE IN ('D', 'D1', 'D2') and CLASSROSTER_CAMP_CODE='1' and (ADDRESSLIST_to_date is NULL OR ADDRESSLIST_TO_DATE > SYSDATE) ) SELECT STUINFO_id , STUINFO_last_name , STUINFO_first_name , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'PR' THEN ADDRESSLIST_STREET_LINE1 END) AS PR_ADDRESSLIST_STREET_LINE1 , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'PR' THEN ADDRESSLIST_STREET_LINE2 END) AS PR_ADDRESSLIST_STREET_LINE2 , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'PR' THEN ADDRESSLIST_city END) AS PR_ADDRESSLIST_city , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'PR' THEN ADDRESSLIST_stat_code END) AS PR_ADDRESSLIST_stat_code , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'PR' THEN ADDRESSLIST_zip END) AS PR_ADDRESSLIST_zip , MIN (CASE WHEN ADDRESSLIST_natn_code IS NULL THEN ADDRESSLIST_natn_code END) AS PR_ADDRESSLIST_natn_code , MIN(NATION_nation) NATION_nation , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'CA' THEN ADDRESSLIST_STREET_LINE1 END) AS CA_ADDRESSLIST_STREET_LINE1 , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'CA' THEN ADDRESSLIST_STREET_LINE2 END) AS CA_ADDRESSLIST_STREET_LINE2 , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'CA' THEN ADDRESSLIST_city END) AS CA_ADDRESSLIST_city , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'CA' THEN ADDRESSLIST_stat_code END) AS CA_ADDRESSLIST_stat_code , MIN (CASE WHEN ADDRESSLIST_atyp_code = 'CA' THEN ADDRESSLIST_zip END) AS CA_ADDRESSLIST_zip FROM got_rnum WHERE rnum = 1 GROUP BY STUINFO_last_name , STUINFO_first_name , STUINFO_id ORDER BY STUINFO_last_name;Max
-
Outer join or a function of the level line
Hi all
I have a question which involves five tables T1, R1, R2, R3, and R4. T1 is an operating table, while R1, R2, R3, and R4 are tables of references (tables parent, with the foreign keys defined in T1). Table T1 contains always referenced data R2 and R1. BUT table T1 can sometimes contain NULL for R3 and R4.
Now my question is simple;
Should I use an OUTER Join for R3 and R4 in the query? as
< code >
Select T1.col1, R1.col2, R2.col2, R3.col2, R4.col2
T1, R1, R2, R3, R4
where T1.col2 = R1.col1
and T1.col3 = R2.col1
and T1.col4 = R3.col1 (+)
and T1.col5 = R4.col1 (+)
< code >
OR
Can I use level functions online for R3 and R4, as
< code >
Select T1.col1, R1.col2, R2.col2,
(Select R3.col2 in R3 where R3.col1 = T1.col4),
(Select R4.col2 in R4 where R4.col1 = T1.col5)
T1, R1, R2
where T1.col2 = R1.col1
and T1.col3 = R2.col1
< code >
Which approach is better and why?
Records in T1 = 2 000 000
R1 = 1000 records
Records in R2 = 300
Records in R3 = 1800
Records in R4 = 200
Please note that all foreign keys are indexed, there are primary keys for all the tables in R
Thank you
QQ.According to the 'official' documentation the outer join should be the by far best choice, because I still think the Oracle documentation says that the subquery will be executed for each row of the main query.
But the Oracle execution engine is a lot smarter (I think from Oracle 8i) and offers a "Value for the filter" feature supposedly which also applies to the "scalar subqueries" you are talking about to (but only if the subquery is considered to be deterministic, i.e. returns always the same value for the same input value of the main request).
Basically, this Treaty optimization the subquery as a function of an input value (the link to the main request) and an output value (the result of the query) and maintains a hash table in memory of pairs of input/output value. The size of the changes table in memory from one version to the other, I think in 9i, it is set by default to 256 entries but in 10g, it is limited in size and therefore the number of entries depends on the size of the values to store.
The effectiveness of this optimization of memory-search in the table depends on certain factors such as the number of pairs of distinct value, the order of the incoming values and collisions in the table which can be occur according to hash values.
Jonathan Lewis has an in-depth coverage of these mechanism in its "cost-based Oracle: Fundamentals" book.
To make a long story short, since your tables R3 and R4 are relatively low, you might want to give a chance, because these optimizations could work very well in your particular case, but unfortunately the real result is simply unpredictable due to the above constraints, including the order of the incoming values.
Kind regards
RandolfOracle related blog stuff:
http://Oracle-Randolf.blogspot.com/SQLTools ++ for Oracle (Open source Oracle GUI for Windows):
http://www.sqltools-plusplus.org:7676 /.
http://sourceforge.NET/projects/SQLT-pp/ -
Bug with an outer join, or & Analytics function (or rownum)
Hello
Seems to be a combination of an outer join, OR and rownum confuses the CBO.
First request is without rownum, the second is with rownum.
The second query expects 203 t lines and never ends. It should behave the same as query 1, with 24 M lines.
Remove the GOLD clause query 2 allows him to behave as a query 1, with 24 M lines.
We never saw it? Is there a solution?
SELECT * FROM message i LEFT JOIN (SELECT hi.message_id, hi.update_dt FROM message_hist hi) h ON (t.id = h.master_id AND(t.update_dt = h.update_dt OR h.update_dt <TO_DATE('150901','RRMMDD'))); ----------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ----------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 24M| 13G| 475G (2)|999:59:59 | | 1 | NESTED LOOPS OUTER | | 24M| 13G| 475G (2)|999:59:59 | | 2 | TABLE ACCESS FULL | MESSAGE | 8037K| 1318M| 29883 (2)| 00:06:59 | | 3 | VIEW | | 3 | 1302 | 59136 (2)| 00:13:48 | |* 4 | TABLE ACCESS FULL| MESSAGE_HIST | 3 | 168 | 59136 (2)| 00:13:48 | ----------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 4 - filter("I"."MESSAGE_ID"="HI"."MESSAGE_ID" AND ("HI"."UPDATE_DT"<TO_DATE('150901','RRMMDD') OR "I"."UPDATE_DT"="HI"."UPDATE_DT")) ---------------- SELECT * FROM message i LEFT JOIN (SELECT hi.message_id, hi.update_dt , ROWNUM FROM message_hist hi) h ON (t.id = h.master_id AND(t.update_dt = h.update_dt OR h.update_dt <TO_DATE('150901','RRMMDD'))); ------------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ------------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 203T| 112P| 475G (2)|999:59:59 | | 1 | NESTED LOOPS OUTER | | 203T| 112P| 475G (2)|999:59:59 | | 2 | TABLE ACCESS FULL | MESSAGE | 8037K| 1318M| 29883 (2)| 00:06:59 | | 3 | VIEW | | 25M| 10G| 59151 (2)| 00:13:49 | |* 4 | VIEW | | 25M| 10G| 59151 (2)| 00:13:49 | | 5 | COUNT | | | | | | | 6 | TABLE ACCESS FULL| MESSAGE_HIST | 25M| 1355M| 59151 (2)| 00:13:49 | ------------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 4 - filter("I"."MESSAGE_ID"="H"."MESSAGE_ID" AND ("I"."UPDATE_DT"="H"."UPDATE_DT" OR "H"."UPDATE_DT"<TO_DATE('150901','RRMMDD')))RowNum in a subquery is supposed to ensure that the subquery is evaluated completely before filtering, otherwise, how could you go out rownum?
Your question is compounded because of the join condition that forces a level of nested loops, which means that the table should be fully analysed once for each line of conduct rowsource. You can either transform the join in an equijoin and allow a hash to run, or you join could materialize the subquery once.
Allow the hash join:
SELECT count (*)
Message FROM
LEFT JOIN (SELECT hi.message_id, hi.update_dt
ROWNUM
OF message_hist salvation) PM ON (i.message_id = h.message_id
AND i.update_dt = h.update_dt)
LEFT JOIN (SELECT hi.message_id, hi.update_dt
ROWNUM
OF message_hist salvation) h2 ON (i.message_id = h2.message_id
AND h2.update_dt<>
AND h2.update_dt <> i.update_dt)
/
----------------------------------------------------------------------------------------
| ID | Operation | Name | Lines | Bytes | Cost (% CPU). Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1. 66. 211 (1) | 00:00:01 |
| 1. GLOBAL TRI | | 1. 66. | |
|* 2 | EXTERNAL RIGHT HASH JOIN | | 800 | 52800 | 211 (1) | 00:00:01 |
|* 3 | VIEW | | 1. 22. 70 (0) | 00:00:01 |
| 4. COUNTY | | | | | |
| 5. TABLE ACCESS FULL | MESSAGE_HIST | 1. 22. 70 (0) | 00:00:01 |
|* 6 | EXTERNAL RIGHT HASH JOIN | | 800 | 35200. 141 (1) | 00:00:01 |
| 7. VIEW | | 1. 22. 70 (0) | 00:00:01 |
| 8. COUNTY | | | | | |
| 9. TABLE ACCESS FULL | MESSAGE_HIST | 1. 22. 70 (0) | 00:00:01 |
| 10. TABLE ACCESS FULL | MESSAGE | 800 | 17600 | 70 (0) | 00:00:01 |
----------------------------------------------------------------------------------------
Information of predicates (identified by the operation identity card):
---------------------------------------------------
2 - access("I".") MESSAGE_ID '= 'H2'.' MESSAGE_ID "(+))"
filter ("H2". "UPDATE_DT" (+)<>'I'. " ("' UPDATE_DT")
3 - filter("H2".") UPDATE_DT "(+)<>
6 - access("I".") "UPDATE_DT" ="H" UPDATE_DT "(+) AND"
"I"." ' MESSAGE_ID ' ="H" MESSAGE_ID "(+))"
Materialize the subquery:
WITH h AS (SELECT / * + MATERIALIZE * / hi.message_id, hi.update_dt)
ROWNUM
OF message_hist salvation)
SELECT count (*)
Message FROM
LEFT JOIN: ON (i.message_id = h.message_id
AND (i.update_dt = h.update_dt OR h.update_dt<>
----------------------------------------------------------------------------------------------------------
| ID | Operation | Name | Lines | Bytes | Cost (% CPU). Time |
----------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1. 22. 1740 (0) | 00:00:01 |
| 1. TRANSFORMATION OF THE TEMPORARY TABLE. | | | | |
| 2. LOAD SELECT ACE | SYS_TEMP_0FD9D6810_5B8F6E67 | | | | |
| 3. COUNT | | | | | |
| 4. TABLE ACCESS FULL | MESSAGE_HIST | 1. 22. 70 (0) | 00:00:01 |
| 5. GLOBAL TRI | | 1. 22. | |
| 6. NESTED EXTERNAL LOOPS | | 800 | 17600 | 1670 (0) | 00:00:01 |
| 7. TABLE ACCESS FULL | MESSAGE | 800 | 17600 | 70 (0) | 00:00:01 |
| 8. VIEW | | 1. | 2 (0) | 00:00:01 |
|* 9 | VIEW | | 1. 22. 2 (0) | 00:00:01 |
| 10. TABLE ACCESS FULL | SYS_TEMP_0FD9D6810_5B8F6E67 | 1. 22. 2 (0) | 00:00:01 |
----------------------------------------------------------------------------------------------------------
Information of predicates (identified by the operation identity card):
---------------------------------------------------
9 - filter("I".") ' MESSAGE_ID ' ="H" MESSAGE_ID' AND ('I'. "" "UPDATE_DT"="H" UPDATE_DT' OR
"H"." UPDATE_DT ".<>
You may need to change the first condition to make sure that you select the correct subquery.
-edit
Not able to view a plan but you can invade the second join condition select and then the result of a subquery with a predicate according to your requirement. This should delay the or rating and leave only an equijoin (although rowsource return may be slightly larger than the opposite).
-Second edition, it did not work exactly when I tried it.
A hybrid of the previous two plans with a slight modification of how he was imitating the GOLD:
WITH h AS (SELECT / * + MATERIALIZE * / hi.message_id, hi.update_dt)
ROWNUM Clotilde
OF message_hist salvation)
SELECT i.MESSAGE_ID
i.UPDATE_DT
COALESCE(h.message_id,h2.message_id) message_id
, COALESCE (h.update_dt, h2.update_dt) update_dt
Clotilde COALESCE (h.rown, h2.rown)
Message FROM
LEFT JOIN: ON (i.message_id = h.message_id
AND i.update_dt = h.update_dt)
LEFT JOIN: h2 WE (DECODE(h.message_id,,i.message_id) = h2.message_id - only try this if previous join returned NULL
AND h2.update_dt<>
/
--------------------------------------------------------------------------------------------------------
| ID | Operation | Name | Lines | Bytes | Cost (% CPU). Time |
--------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1. 66. 8 (0) | 00:00:01 |
| 1. TRANSFORMATION OF THE TEMPORARY TABLE. | | | | |
| 2. LOAD SELECT ACE | SYS_TEMP_0FD9D6605_28F27F | | | | |
| 3. COUNT | | | | | |
| 4. TABLE ACCESS FULL | MESSAGE_HIST | 150. 3300 | 2 (0) | 00:00:01 |
| 5. GLOBAL TRI | | 1. 66. | |
|* 6 | EXTERNAL RIGHT HASH JOIN | | 10497. 676K | 6 (0). 00:00:01 |
|* 7 | VIEW | | 150. 3300 | 2 (0) | 00:00:01 |
| 8. TABLE ACCESS FULL | SYS_TEMP_0FD9D6605_28F27F | 150. 3300 | 2 (0) | 00:00:01 |
|* 9 | OUTER HASH JOIN | | 328. 14432 | 4 (0) | 00:00:01 |
| 10. TABLE ACCESS FULL | MESSAGE | 200 | 4400 | 2 (0) | 00:00:01 |
| 11. VIEW | | 150. 3300 | 2 (0) | 00:00:01 |
| 12. TABLE ACCESS FULL | SYS_TEMP_0FD9D6605_28F27F | 150. 3300 | 2 (0) | 00:00:01 |
--------------------------------------------------------------------------------------------------------Information of predicates (identified by the operation identity card):
---------------------------------------------------6 - access("H2".") MESSAGE_ID "(+) = DECODE (TO_CHAR ('H'". "))" MESSAGE_ID"), NULL, 'I '. (("' MESSAGE_ID '))
7 - filter("H2".") UPDATE_DT "(+)<>
9 - access("I".") "UPDATE_DT" ="H" UPDATE_DT "(+) AND 'I'." "" ' MESSAGE_ID '="H" MESSAGE_ID "(+))"(This plan is another system if costs are not comparable)
-
Oracle: Use LEFT OUTER JOIN, but convert the data to an external list
Hello, all,.
I know it can be done; I just don't remember how I got it done, oh there are so many years.
Assumes that the tables exist for groups and individuals. People can belong to several groups.
SELECT g.groupName, p.lastName || ', ' || p.firstName as fullName FROM groups g LEFT OUTER JOIN groupPersonAssociation gpa ON gpa.groupID = g.groupID LEFT OUTER JOIN person p ON p.personID = gpa.personID ORDER BY g.groupName, fullName
This gives us:
Group One Alpha, Daniel Group One Bravo, Charles Group One Charlie, Chuck Group Two Beta, Alpha Group Two Delta, Bonnie Group Three Echo, Bunny Group Three Golf, Samuel Group Three November, Stan
How word the SQL to get the data as:
Group One Alpha, Daniel | Bravo, Charles | Charlie, Chuck Group Two Beta, Alpha | Delta, Bonnie Group Three Echo, Bunny | Golf, Samuel | November, Stan
V/r,
^_^
I finally thought to it. I was using incorrect keywords on Google.
SELECT g.groupName, LISTAGG(p.lastName || ', ' || p.firstName,' | ') WITHIN GROUP (ORDER BY g.groupName) "fullName" FROM groups g LEFT OUTER JOIN groupPersonAssociation gpa ON ggpa.groupID = g.groupID LEFT OUTER JOIN person p ON p.personID = gpa.personID GROUP BY g.groupName ORDER BY g.groupName, fullNameJust in case someone else is going through this same desire.
HTH,
^_^
-
Hi all
I use under version
Connected to Oracle Database 11g Express Edition Release 11.2.0.2.0
SQL > SELECT E.ENAME,.
2 D.DEPTNO,
3 D.LOC
4. TO EMP E,.
DEPT 5 D
6. WHERE = E.DEPTNO D.DEPTNO (+);
ENAME, DEPTNO LOC
---------- ------ -------------
DALLAS SMITH 20
ALLEN 30 CHICAGO
WARD 30 CHICAGO
20 DALLAS JONES
MARTIN 30 CHICAGO
BLAKE 30 CHICAGO
CLARK 10 NEW YORK
SCOTT 20 DALLAS
THE 10 NEW YORK KING
TURNER 30 CHICAGO
20 DALLAS ADAMS
JAMES 30 CHICAGO
FORD 20 DALLAS
MILLER 10 NEW YORK
40 BOSTON
15 selected lines
-----------------------------------------------------------------------------------------------------------------------------
SQL > SELECT E.ENAME,.
2 D.DEPTNO,
3 D.LOC
4. TO EMP E
5 LEFT OUTER JOIN
D 6 DEPT
7. THE E.DEPTNO = D.DEPTNO;
ENAME, DEPTNO LOC
---------- ------ -------------
MILLER 10 NEW YORK
THE 10 NEW YORK KING
CLARK 10 NEW YORK
FORD 20 DALLAS
20 DALLAS ADAMS
SCOTT 20 DALLAS
20 DALLAS JONES
DALLAS SMITH 20
JAMES 30 CHICAGO
TURNER 30 CHICAGO
BLAKE 30 CHICAGO
MARTIN 30 CHICAGO
WARD 30 CHICAGO
ALLEN 30 CHICAGO
14 selected lines
My doubt is both are same query is the same, is in ansi format and is in the format of the Oracle,.
but the results are different.
For the first query null is coming for unmatched records in the dept table
but in the second query, it does not come
Thank you
Hello
2947022 wrote:
Hi all
I use under version
Connected to Oracle Database 11g Express Edition Release 11.2.0.2.0
SQL > SELECT E.ENAME,.
2 D.DEPTNO,
3 D.LOC
4. TO EMP E,.
DEPT 5 D
6. WHERE = E.DEPTNO D.DEPTNO (+);
ENAME, DEPTNO LOC
---------- ------ -------------
DALLAS SMITH 20
ALLEN 30 CHICAGO
WARD 30 CHICAGO
20 DALLAS JONES
MARTIN 30 CHICAGO
BLAKE 30 CHICAGO
CLARK 10 NEW YORK
SCOTT 20 DALLAS
THE 10 NEW YORK KING
TURNER 30 CHICAGO
20 DALLAS ADAMS
JAMES 30 CHICAGO
FORD 20 DALLAS
MILLER 10 NEW YORK
40 BOSTON
15 selected lines
-----------------------------------------------------------------------------------------------------------------------------
SQL > SELECT E.ENAME,.
2 D.DEPTNO,
3 D.LOC
4. TO EMP E
5 LEFT OUTER JOIN
D 6 DEPT
7. THE E.DEPTNO = D.DEPTNO;
ENAME, DEPTNO LOC
---------- ------ -------------
MILLER 10 NEW YORK
THE 10 NEW YORK KING
CLARK 10 NEW YORK
FORD 20 DALLAS
20 DALLAS ADAMS
SCOTT 20 DALLAS
20 DALLAS JONES
DALLAS SMITH 20
JAMES 30 CHICAGO
TURNER 30 CHICAGO
BLAKE 30 CHICAGO
MARTIN 30 CHICAGO
WARD 30 CHICAGO
ALLEN 30 CHICAGO
14 selected lines
My doubt is both are same query is the same, is in ansi format and is in the format of the Oracle,.
but the results are different.
For the first query null is coming for unmatched records in the dept table
but in the second query, it does not come
Thank you
In fact, these requests are not the same.
The first is to find all the lines of the Department, with the corresponding lines of PGE (when there are). This is equivalent to «FROM dept LEFT OUTER JOIN emp...» ».
The second is to find all the rows in the emp of the lines of the Department (when there are any). This is equivalent to «...» WHERE e.deptno = d.deptno (+).
-
Hi gurus,
Left outer join:
-------------------
Select * from A LEFT OUTER JOIN B on A.col = B.col;
by definition, a left outer join brings the rows that match the join condition and lines not corresponding to table A.
My question here is "corresponding to B and no matching rows in A" is that nothing, but... SELECT * FROM A;
can someone pls clarity...
Thank you.
Imagine that you had:
TableA
COLUMN1 COLUMN2
'A' 1
'B' 2
'C' 3
TABLEB
COLUMN1 COLUMN2
'A' 'X1'
'A' 'X2'
'B' 'Y'
'D' 'Z'
SELECT * FROM TABLEA;
A 1
B 2
C 3
Now, if you want to join (first column is either in table A or B, (deuxieme from tableA, third table B)
SELECT * FROM TABLEA A JOIN TABLEB B ON (A.COLUMN1 = B.COLUMN1)
A 1 X 1
A 1 X 2
B 2 Y
SELECT * FROM TABLE LEFT B EXTERNAL ON (A.COLUMN1 = B.COLUMN1) JOIN TABLE
A 1 X 1
A 1 X 2
B 2 Y
C 3 {null}
SELECT * FROM TABLE A TABLE RIGHT OUTER JOIN B (A.COLUMN1 = B.COLUMN1)
A 1 X 1
A 1 X 2
B 2 Y
D {null} Z
SELECT * FROM TABLE A TABLE FULL OUTER JOIN B (A.COLUMN1 = B.COLUMN1)
A 1 X 1
A 1 X 2
B 2 Y
C 3 {null}
D {null} Z
HTH
-
Bad result in a left outer join in 12.1.0.2
Hallo,
We discovered a strange behaviour in a query. The query provides values in a column of outer join where there is no corresponding value in the table is attached to the outside.
When you expand this request by the "ORDER BY" then this query gives the correct result.
Example:
SQL > desc tb_a
Name Null? Typ
-------------------------------------------- ----------------------------
ID NOT NULL NUMBER (19)SQL > desc tb_b
Name Null? Typ
-------------------------------------------- ----------------------------
CLOSED NOT NULL NUMBER (1)
ID NOT NULL NUMBER (19)CCS_APPLICATION@icw01> select * from tb_a where id in (4148,4141,4195);
ID
----------
4148
4141
4195CCS_APPLICATION@icw01> select * from tb_b where id in (4148,4141,4195);
INTERNAL ID
---------- ----------
4148 0CCS_APPLICATION@icw01> SELECT
2 b.id AS b_id,
3 a.id AS a_id,
4 b.closed AS b_closed
5
6 tb_a a
7 LEFT OUTER JOIN tb_b b ON a.id = b.id
8 WHERE a.id IN (4148, 4195, 4141)
9 ORDER BY ASC a.id
10;B_ID ALLOCATION A_ID B_CLOSED
---------- ---------- ----------
4141
4148 4148 0
4195CCS_APPLICATION@icw01> SELECT
2 b.id AS b_id,
3 a.id AS a_id,
4 b.closed AS b_closed
5
6 tb_a a
7 LEFT OUTER JOIN tb_b b ON a.id = b.id
8 WHERE a.id IN (4148, 4195, 4141)
9 - ORDER BY ASC a.id
10;B_ID ALLOCATION A_ID B_CLOSED
---------- ---------- ----------
4148 4148 0
4141 4141
4195 4195instance parameter:
VALUE OF TYPE NAME
------------------------------------ ----------- ------------------------------
compatible string 12.1.0.2.0
optimizer_features_enable string 12.1.0.2
After ""alter system set optimizer_features_enable = ' 11.2.0.4 ';" the query provides the correct result in both cases (ordered and unordered).
Now the final question: is this a bug?
1480970 wrote:
Hallo! Yes, I searched the Support of Oracle. I found some similar entries, but not an exact match. To fix some issues
with 12.1.0.2.
There is another interesting clue when look you on the execution plan:
Note
-----
-the dynamic statistics used: dynamic sampling (level = 2)
- This is an adaptation plan
We have disabled (= FALSE) optimizer_adaptive_features and the query provides the correct values.
This could be a solution for us.
Looks like a pretty tight match for bug 18430870, even if it affects the two 12.1.0.1 and 12.1.0.2, which contradicts the Martin trial against 12.1.0.1.
The description of the bug mentions disabling "_projection_pushdown" (set to false) should also be a viable solution, perhaps if you want to give that a try and see if it is a different bug or not.
There are also a number of one-time fixes already available for download, maybe your version / platform is already covered, if the bug applies.
Randolf
-
Hi all
I have 3 tables A, B, C
Create table a (varchar2 (100)) of the currency;
insert into a values (GBp);
insert into a values (GBP);
insert into a values (GBX);
Create table B (varchar2 (100) currency, number);
insert into B values (GBP, 61.1);
Create a table (minor_currency varchar2 (100), major_currency varchar2 (100));
insert into values of C (GBp, GBP);
insert into values of C (GBX, GBP);
I need to get the rate table B by linking the A with the currency as a condition of joining. (left outer join)
For currencies which are not in table B, table should be attached with C minor currency-based
and get the major_currency and join with table B
Ex:
something like this:
Select B.rate from A, B, C
WHERE (A.currency = B.currency or (A.currency = C.minor_currency and B.currency = C.major_currency)
O/P: for GBp and GBX currency, I need to get the rate as 61.1 in table B, but B currency is GBP. So I need to get the major_currecny for GBp, GBX table C and join with the table B
Thank you
Sasi
Hi all
Thanks for your time. Its done
Maybe you are looking for
-
'On this day' feature for pictures of mac, it exists?
Hello I like the feature of Facebook "on this day" but would like to have on my mac instead of downloading my collection of entire photo on FB for this to work over one year, etc.. Does anyone know how to reproduce it in a script, 3rd party app etc.?
-
Equium A60-181 screen is very low after reconnection of cable
My satellite had been very slow lately and I was advised to remove the base and clean the dust fan (I'm quite a novice). When I replaced the screw, the screen would not work, but all the power lights were on. I reopened the base and found that a wire
-
Unable to save the files in Windows Movie Maker
NEED HELP now Windows Movie Maker CRASHING! Project due tomorrow Help! It is 9: 40 EST and my video girls ' school project is due tomorrow, my WMM crashes when I try to "Save movie file" is only 6 minutes long and time continues to go up instead of d
-
Restored Windows 8.1 and I lost email storage folders - can someone advise
10 Windows installed ok and I was happy - but some 3rd party software programs not functioning properly - tried some reinstalls and messed up settings on their opinion correct - therefore restored to windows 8.1 and have lost all my email storage fil
-
Disable the window of creating Adobe PDF permanently
When I create a PDF document, a small window will appear called "create Adobe PDF", which shows the progress. Is generally visible for a few seconds. Is it possible to disable this window given permanently?