SQL & PL/SQL :: Selecting First Row From Multiple Joins?
Apr 2, 2010
I'm putting together a path to select a revision of a particular novel:
SELECT e.documentname, e.Revision, e.VersionNumber
FROM Catalog, BookInCatalog
INNER JOIN NovelMaster
INNER JOIN HasNovelRevision
INNER JOIN NovelRevision e
LEFT JOIN NovelRevision s
[code]...
My goal here is to select the earliest revision from the set of Novel revision. The revision field is a string.
When I run the query for Novels that have multiple revisions I get multiple records. If there is just one record I only get one row. If there are two I get four (two for each revision). As the number of revision increases it looks like it just mushrooms from there.
One other challenge is the format of the revision- a revision sequence could look like this:
A
B
C1
C2
C
D
E1
E
So there are "intermediate" revision referred to by a number. In this case I would select revision A, but if I had:
A1
A
B1
B
I would want to select B. I am pretty sure that all the revision are stored in the db in order.Notice that the comparison operator ">" is used in e.Revision > s.Revision. I initially though it should have been "<" because we want to select the initial but the other way gives me the right order (though the wrong results).
I need to calculate a list of people, who got some services more that 2 times with the same service koda (pas_kodas) to the same person (zmo_kodas). It should not depend on report number.
[URL]...
What I get is in green (services are calculated more than 2 times BUT in the same report).
What I need is in red: calculate servises more that 2 times ACCROSS all reports to the same person (zmo_kodas).
[URL]...
One person (zmo_kodas) can have a lot of reports (ats_nr).
Every report can have one or more services (pas_kodas).
I need to calculate a list of people, who got some services more that 2 times with the same service koda (pas_kodas) to the same person (zmo_kodas). It should not depend on report number.
[URL]
What I get is in green (services are calculated more than 2 times BUT in the same report). What I need is in red: calculate servises more that 2 times ACCROSS all reports to the same person (zmo_kodas).
[URL]
One person (zmo_kodas) can have a lot of reports (ats_nr). Every report can have one or more services (pas_kodas).
The structure is like this may contain multiple records like Comp_id, Comp_name, ISIN will be same, but column_name will contain the column_name to which its corresponding column_value needs to be populated to.
I am working on Pro*C and i have a requirement where i need to select all the rows from a table into a c - structure variable. Since i get to know the no of rows in the table which is getting selected only at run time, i need to create a pointer variable to the structure and i'll allocate the size to it based on the count of rows in the table using malloc or calloc.I tried allocating memory using calloc and it does not show any error. But when i when the exec select statement run it shows an error.
Statements i have used: struct common *comp; struct common_ind *comp_i;
We have to load 10 million rows in a table from another table based on the multiple joins. How much tablespace size we allocate to the table and for performance point of view how much should be the SGA size.
I have a table which contains the multiple records for single ID No. Now i have to select single record which contains the latest date. here is the structure Name
Null Type ------ ---- ------------ ID_P NUMBER NAME_P VARCHAR2(12) DATE_P TIMESTAMP(6) Records---------------------1 loosi 22-AUG-13 01.27.48.000000000 PM1 nammi 26-AUG-13 01.28.10.000000000 PM2 kk 22-AUG-13 01.28.26.000000000 PM2 thej 26-AUG-13 01.28.42.000000000 PM
now i have to select below 2 rows how can write select qurie for this?
its a very basic query, i want to open multiple .fmb / .rdf files at once ( like pressing Shift or Ctrl key when selecting multiple files ), but its not allowing me
In mys store procedure I am using a subquery with INNER JOIN. This subquery calls a user defined function which takes main query fields as parameter values. But i am facing issue for accessing main query fields. my query is like below:
SELECT ID, Name, sub.Desc, sub.Date FROM MainTable main INNER JOIN ( SELECT * FROM RMF_GetData(main.ID) ) sub ON main.ContactID = sub.ContactID
I need data from a function in a table format based on some case conditions. Hence i need to join it with main table. But here oracle gives error as "invalid identifier" for main.ID parameter.
SELECT so.* FROM shipping_order so WHERE (so.submitter = 20) OR (so.requestor_id IN (SELECT poc.objid FROM point_of_contact poc WHERE poc.ain = 20)) OR so.objid IN (SELECT ats.shipping_order_id FROM ac_to_so ats WHERE (ats.access_control_id IN (selectac. objid FROM access_control ac WHERE ac.ain = 20 OR ac.group IN ('buyers', 'managers'))))
rewrite this query to use joins. That would greatly simplify my sql query building code. The ids, objids, submitter, ain are numeric and group is a varchar.
I pulled in 1121 SSN's into a table and am using that table as the basis for returning data from other tables...including how many documents a user has in their folder.
My query; however, is only returning 655 rows...it is returning only those rows that have documents in their folders. I want to return ALL rows...WHETHER OR NOT THEY HAVE A DOCUMENT COUNT (count(*)). How can I get all 1,121 rows to return? I would like the output to look like:
SSN LOCATION EMP_STATUS FOL_STATUS COUNT(*) -- For those folders containing documents: XXX-XX-XXXX WHATEVER WHATEVER WHATEVER 12
-- For those folders containing 0 documents: XXX-XX-XXXX WHATEVER WHATEVER WHATEVER NULL
Here is the query in it's current state:
-- Get User/Folder/Doc Count Information SELECT b.ssn, b.location, b.emp_status, c.fol_status, COUNT (*)
[Code]....
So again, my problem here is that...not all FOLDERS contain DOCUMENTS...but I still want the folder data lised...I just need it listed with either a zero count (0), or a NULL in the COUNT(*) column.
I'm trying the various joins, but none of them seem to be working.
I've tried the old 8i (+) join as follows:
AND c.fol_id = d.doc_fol_id(+)
I've tried the inner join:
AND c.fol_id(+) = d.doc_fol_id
...and I've tried the 9i method (left outer and full outer) using the following types of notations:
folder c full outer join documend d on c.fol_id = d.doc_fol_id
...so far, no luck. I'm still having only 655 rows returned (the 655 are those folders that HAVE document count > 0. Any folder that has zero documents in the document table just aren't being returned in the query.)
why how ever way i try i cant get the joins on the tables properly.... well i know i have to work hard....if join is not proper the data i extract is also not proper.Well now i have 3 tables...
ps_operations
Name Null? Type ----------------------------------------- -------- -------- ASSB_PT_NBR_SEQ_ID NOT NULL NUMBER(8) OPERATION_NBR NOT NULL VARCHAR2(10) EFFECTIVE_FROM_DT NOT NULL DATE DML_TS NOT NULL DATE DML_USER_ID NOT NULL VARCHAR2(30) OPERATION_DESC NOT NULL VARCHAR2(70) HOURS_PER_PIECE_QTY NOT NULL NUMBER(9,6) PIECES_PER_HOUR_RATE_QTY NOT NULL NUMBER(15,7) EFFECTIVE_TO_DT DATE EXTRACT_IND VARCHAR2(1) [code]...
I have never worked on CPK and UK....so i dont know how to use them to join the tables,.
We have access to a remote Oracle database in Germany and need to insert selected data to our local Oracle database. Our problem is that we have to do several joins (7 tables) on the remote database as well as using one where clause (always the same: P.T_LIEF_LFNT_1='12803193').
Unfortunately we do not have rights to create a view on the remote database. Is there another way to send the entire query to the remote database for processing? Also, the query returns approximately 34,000 rows.
Here is our current query:
INSERT INTO PRIMUS(PARTNO, SORT_FORMAT, TP_WORKSPACE,
I have a query I am trying to tune. It presently takes anywhere from 15 minutes to two hours to run, depending on how many records the client has. But it needs to run several hundred times, and finish over the course of a weekend. When it runs over, we have problems.
Here's the basic structure of the query:
CODESELECT ... FROM main_tab, tab_a, tab_b, tab_c, ... tab_z WHERE main_tab.client_id = :1 AND main_tab.unique_id = tab_a.unique_id(+) AND main_tab.unique_id = tab_b.unique_id(+) AND main_tab.unique_id = tab_c.unique_id(+) ... AND main_tab.unique_id = tab_z.unique_id(+);
All of the tables are indexed (and statistics are gathered) on the field unique_idMain_tab has an index on client_id.There is a one-to-one join (sometimes one-to-zero, thus the outer join) from the main_tab table to all the other tables.These are static tables, they're wiped and recreated - no changes, inserts, deletes.
By default, the optimizer does a full table scan and then a hash join on every single of the 26 tab_a through tab_z tables, only using the index on main_tab.
By the way I can add indexes, possibly even to the point of adding an index on some tables that would include all the fields found in the select clause on that table. But I cannot change the table structure (by, say, combining these tables together).
To explain, if first value is null, then TBL_EQUIPANDSERVICES.LastWOSeq = tbl_workorders.workorderseq (+), otherwise join the transaction.workorderseq.
When I try and execute this code, I'm given "ORA-01417: a table may be outer joined to at most one other table." I've double checked, TBL_Workorders is not joined with any other table in any select.
Now, the EquipandServices.workorderseq is joined (+) with another table.
I am working on a new project in OBIEE. I am asked to do the data modeling in the database using oracle sql developer. I have to create the joins based on the requirements. I have the tables created already. But the primary keys for few tables are not defined for few tables. PK-FK joins are also not done properly.
My questions are (1) If I have to define the primary key for the existing tables can I do that using the alter table command or should I create the table all over again and then define it? (2) If I have to make the changes in the existing PK-FK joins how do i go about doing that?
I am observing some skewed results for Left outer join where the main table has NULL in the field we are joining against with another table.
Just wondering if there are some tricks to get over it. I am currently using NVL(tab1.col1,'X') = NVL(tab2.col3,'X') and am just wondering if there is a better way to handle this.
merge into merge_st t using (select * from merge_st1) src on (t.id=src.id(+)) when matched then update set name=src.name;
I need to convert oracle joins to ANSI joins. I have tried below
update (select t1.name as t1_name,t2.name as t2_name from merge_st t1 lef outer join merge_st t2 on(t1.id=t2.id)) set t1_name=t2_name;
It statements shows error like cannot modify the non key preserved tables.I have cheked these table has contains whether primary key or not.there is no constraints for these tables.Our application, constaints handle in front end. so we cannot create any constraints in oracel database.how to convert oracle joins to ansi join?
create table test_a (id number, b varchar2 (20)); create table test_b (id number, a number, b number, c number, d number, e number, f number); insert into test_a values (1,'Manu'); insert into test_a values (2,'Tanu'); insert into test_a values (3,'Anu');
[code].....
convert the query above using joins instead of scalar queries, as scalar queries decreasing the performance.
We have two tables, TableA and TableB that contain list of accounts and balances.The requirement is to compare the balances of accounts in both the tables, and if there is a difference, then record that difference with account number in another table.
Both TableA and TableB contain more than 10 million rows.What is the best way to do this task in PL/SQL? A join on TableA and TableB to know the differences has become very slow due to large volume.
i have two tables having some fields .In table table1 there is 2 lakh rows and in table table2 there is 3 lakh rows.In both table 2 lakh rows are common.I have to find those 1 lakh rows which are distinct but the condition for this i have to use only sql joins not minus or subquery or any function.
when i doing this.select t2.id,t2.name,t2.loc from table1 t1,table2 t2 where t1.id <>t2.id;
I'm experiencing some infinite loop for my delete. I tried many way to deal with this problem but still take too much time. I will try to be clear as possible.
I have 4 implicated table in this problem.
The deletion is done depending of the pool_id given
Table 1 contain the pool_id Table 2 the ticket_id foreign join ticket_pool_id with the pool_id Table 3 ticket_child_id foreign join ticket_id with the ticket_id Table 4 ticket_grand_child_id foreign ticket_child_id join with the ticket_child_id
Concerned count for each table 1---->1 table 2---->1 200 000 table 3---->6 300 000 table 4---->6 300 000
So in fact it`s 6.3M+6.3M+1.2M+1 row to be deleted
Here`s the constraint : -No partintionning -Oracle version 9 -Online all the time so no downtime neither CTAS -We cannot use cascade constraint -The normalization is very important
Here`s what I tried: -Bulk delete -Delete with statement (In and Exists clause) -temp table for each level and 1 level join -procedure and commit each 20k
None of those worked in a decent time frame like less then one hour. The fact that we cannot base a delete on one of the column value is not working. Is there a way I'm getting desperate now
