SQL & PL/SQL :: Linking Table Challenge - Showing A Value Only Once Not For Each Match?
May 23, 2012
TABLE 1 : COSTS
I have a table of Costs. We have Jobs that run and there will be a cost associated with a particular machine.So JobNo 1 may run on Machine A and have a cost of 50 dollars. Although its now shown below JobNO 1 could run on MachineB and so on.
We have operators (PERSONCODE) run the jobs on the machines. So Job 1 may be run by PERSONCODE 8 (e.g. Tony) and it may run on MachineA or MachineB.Multipe people may run a particular job. The PERSONCODE will be unique to the Job and it is actually unique to the list. A person never works on more than one job.
Let's say we have Table - A and we would like to replicate specific row transaction to Table B.
Here are the rows in *Table A* Time: Lets say 15:00
A1 Just Updated @15:00 A2 Just inserted @15:01 A3
B1 - Daily Delete Row -i.e just deleted a while back - Non scheduled process --executed by application @15:02 B2 - B3 - Daily Delete Row - i.e just deleted a while back -- Non Schduled process --executed by application @15:05 B4 - Just recently purged (As part of 180 Day purge ) - Scheduled process executed by operations team @15:10 B5 - Just recently purged (As part of 180 Day purge ) - Scheduled process executed by operations team @15:10 B6 -Just recently purged (As part of 180 Day purge ) - Scheduled process executed by operations team @15:10
Current Data in Table B (Before Replication) @15:00
A1 (without updates) A3 B1 B2 B3 B4 B5 B6
Expected rows in Table B (via replication/snapshot/materialized view / or any other method)
*Replication at 15:30* Table B - Read Only
Expected rows after replication-
A1 -- Newly updated details A2 -- Newly inserted row A3 B1 - Daily delete row is expected to be replicated B2 B3 - Daily delete row is expected to be replicated
***Note row B4 is not expected to be replicated to table B.
Questions:
1) How can we get updates, inserts and daily deletes replicated while ignore large purges? 2) How can large purge changes be reflected in replicated tables as well without deleting daily deletes?
There is an interface table and there is an normal transcational table..interface table is being compared with normal table and if match found the result is dumped into another normal table.
I am using two cursors one is to query the interface table and in a for loop pass the results to the second cursor..The interface table is having 5000 + rows and the transcation table is having more than 3.7 millions ..and the program is taking lots of time to execute..took almost 35-45 minutes..
I am using two databases one prd1 and another ftn. I want to do some data verification between these two databases. So I created the following db link in prd1 and also in ftn.
CREATE PUBLIC database LINK ftn CONNECT TO ftn_link IDENTIFIED BY password USING ftn
CREATE PUBLIC database LINK prd1 CONNECT TO prd1_link IDENTIFIED BY password USING prd1
I gave the following sql command from prd1 but it showed error message.
ORA-12154: TNS:could not resolve service name
SELECT A.CUST_ID_MSS_OWNER, B.CUST_ID_MSS_OWNER FROM MSS_V_CAR_REC A, mss_v_car_rec@ftn b WHERE a.lic_no_org = b.lic_no_org;
I have already created the sql net configuration also.
I have two tables as per attachement - TABLE_A has Vehicle details while TABLE_B has Address details:
I am trying to build a query in order to link the Vehicle Details with their Address Details.
I would require an SQL that links records in TABLE_A with records in TABLE_B when group_number are the same; however links with group_number '999' (for the same customer_number) if there are no exact matches.
SELECT A.*,B.* FROM TABKE_A A, TABLE_B B WHERE ((A.CUSTOMER_NUMBER = B.CUSTOMER_NUMBER AND A.GROUP_NUMBER = B.GROUP_NUMBER) OR (A.CUSTOMER_NUMBER = B.CUSTOMER_NUMBER AND A.GROUP_NUMBER = '999'))
The only problem with such query is that record in TABLE_A with group_number '456' will return 2 times. One with address having group_number '456' and one with group_number '999'.
Once I add an extra column to link the two tables the number of rows is reduced by one, however when I try to get this extra record I find none matching the same criteria that led to reducing the records.
SQL> SELECT count(*) 2 FROM N_CONTRACT NC, N_WITHDRAWAL_REQUEST NWR 3 WHERE NC.fk_temP_withdrawal_req_serial = NWR.SERIAL_NUMBER; COUNT(*) ---------- 2243 [code]...
I remember a method in which I used a property in the property pallete to specify the query in which i wanted the list linked with, it gave me the option to input a static list or as mentioned a dynamic list which was dependent on a query entry, but I cant for the life of me remember where this property was.
I am trying to build a report.My query is working fine when i take out this report for a single area_code.But it is not showing proper result when report is take for all are_code's available in table.I have used two tables transactions and balance
create table transactions ( glcode varchar2(10), area_code varchar2(10), debit number,credit number ); insert into transactions values(2000,'ap',200,200); insert into transactions values(3000,'ap',222,222); insert into transactions values(4000,'ap',123,123); insert into transactions values(2000,'dp',200,200); insert into transactions values(3000,'dp',222,222); insert into transactions values(4000,'dp',123,123); insert into transactions values(2000,'pp',200,200); insert into transactions values(3000,'pp',222,222); insert into transactions values(4000,'pp',123,123); [code]....
Suppose I have a table in which I have first_name, last_name, dob. Now I have to fetch on the basis of first_name=some_value, last_name=some_value and dob=some_date. I want to sort it on the basis of exactly fetched values. Let me take an example-
test table contains-
first_name last_name dob ---------- --------- ---- Manu Batham 02-Feb-1988 Manu Sharma 01-Jul-1987 Avinash Pandey 03-Feb-1988 Ankit Gupta 02-Feb-1988 Manu Aggrawal 02-Feb-1988 Manu Batham 20-Jan-1985 Sikha Batham 17-Apr-1988
first_name last_name dob ---------- --------- ---- Manu Batham 02-Feb-1988 Manu Aggrawal 02-Feb-1988 Manu Batham 20-Jan-1985 Manu Sharma 01-Jul-1987 Ankit Gupta 02-Feb-1988 Sikha Batham 17-Apr-1988
My result is based on the approach- if matched first_name, last_name, dob --> 1st prefrence in order if matched first_name, dob --> 2nd prefrence in order if matched first_name, last_name --> 3rd prefrence in order if matched last_name, dob --> 4th prefrence in order if matched first_name --> 5th prefrence in order if matched last_name --> 6th prefrence in order if matched dob --> 7th prefrence in order
I designed the following query for the same-
Select first_name,last_name,dob,1 "Order" from test Where first_name='Manu' and last_name='Batham' and dob=to_date('02/02/1988','dd/mm/yyyy') union Select a,b,c,2 from test Where [code]......
I know that this is not the best possible solution as the table is very big and doing so many hits on that table will certainly decrease the performance.
Seemingly a very simple problem, but I can't seem to figure it out. Table below is a table that holds UserId' s and LinkId's.
CREATE TABLE UserLink (UserId NUMBER, LinkId NUMBER ); ALTER TABLE UserLink ADD PRIMARY KEY (UserId, LinkId); INSERT INTO UserLink VALUES (1, 1); INSERT INTO UserLink VALUES (1, 2); INSERT INTO UserLink VALUES (3, 1); INSERT INTO UserLink VALUES (4, 1);
I'm looking for a query in which I can plug a List of LinkId's that outputs only users that have a relationship with ALL LinkId's provided. In this example you may assume that the List is equal to (1, 2), but you can't make any assumptions about the size of the list, other than size != 0. The query to find users that match ANY of the provided LinkId's is quite simple:
SELECT UserId FROM UserLink WHERE LinkId IN (1,2)
The correct query should output:
USERID ------ 1
And it should work for any size > 0 list of values.
I have a problem with a query I'm trying to run. I need to match two columns containing names, first column (NAME1) contains only the surname and the second column (NAME2) contains a surname and initials, with the initials turning up on either side of the surname.
create table sample1 ( i number , j date, k number)
insert into sample1 values (1,'23-Apr-2010',11) insert into sample1 values (2,'22-Apr-2010',12) insert into sample1 values (3,'21-Apr-2010',13) insert into sample1 values (4,'19-Apr-2010',14) insert into sample1 values (5,'18-Apr-2010',15) insert into sample1 values (6,'17-Apr-2010',16)
I would like to get nulls , if there is no data for a date. As we can see , here i am missing the data for '20-Apr-2010'.
I did it through "UNIX" , but it's not efficient.
The data might be missing for the complete week also. I need to test in this way only for the last 7 days. I tried something like this:
select i, j , sum(k) from sample1 where j in (select to_date(sysdate - rownum) from dual connect by rownum < = 7) group by i, j
WHERE TO_DATE(TO_CHAR(RR.PEROFOPFROM,'DD-MON-YYYY')||RR.AIRCRAFTSTD,'DD-MON-YYYY:HH24MI') > TO_DATE(TO_CHAR(RR.PEROFOPFROM,'DD-MON-YYYY')||RR.AIRCRAFTSTA,'DD-MON-YYYY:HH24MI')
I have data like this:
PEROFOPFROMAIRCRAFTSTD
29/03/20102150 NULL NULL NULL NULL 30/03/20102150
When I execute the query it always gives me the error "literal does not match format string".
I'm trying to do data mining on a web log which recorded one day web access information from a busy web server. I imported the data into Oracle Data miner, and created a table (WEBLOG). The idea is to create a new field, i.e. session, for the users so that each session could be thought as a representative of a user-intent (aka topic). Now based on this, data mining models would be used to cluster(group) the users based on their similarity. The first step is to prepare the data which involves using SQL queries. So first, all I did was to create a function for date and time. This is the following code I used,
create or replace function ssndate(p_date in varchar2 default '03-01-18', p_time in varchar2) return number $if dbms_db_version.ver_le_10 $then deterministic $elsif dbms_db_version.ver_le_11 $then result_cache $end as begin return trunc((to_date(p_date||' '||p_time, 'dd-mm-yy hh24:mi:ss') - to_date('01-01-90','dd-mm-yy')) * (86400/2400)); end ssndate; /
The function ssndate compiled successfully.The next step I took was to create a view through the following query,
create or replace view WEBLOG_VIEWS as select (select ssndate(LOG_DATE, LOG_TIME) from dual) as "SESSION_DT", C_IP, CS_USER_AGENT, (CS_URI_STEM||'?'||CS_URI_QUERY) as WEB_LINK from WEBLOG;
This was successful as well. The problem is in the next step where I try to do data grouping.
create table FINAL_WEBLOG as select SESSION_DT, C_IP, CS_USER_AGENT, listagg(WEB_LINK, ' ') within group(order by C_IP, CS_USER_AGENT) "WEB_LINKS" from WEBLOG_VIEWS group by C_IP, CS_USER_AGENT, SESSION_DT order by SESSION_DT [code]....
the to_date function should be fine. In the data that I possess, the date and time are in no format. Example: 30118 and 0:00:09 respectively.
insert into iostat_mac values(to_date('2012/03/28 08:00:00 AM'),'653.6','20.7','15392.0','451.8','0.0','5.5','0.0','8.2','0','64','/vol/sun_dc_u2');
Receive ora-01861 error when trying to insert.Suspected it had to do with the date format, I changed the nls parameter for my session to include the time format, but to no avail.
I would like to add an additional column indicator (Y or N) to specify which ID's do not contain all records of 'NONE'. There can be an occurrence of 'NONE' as long as there is another cust_num different to 'NONE' These should be marked as 'Y' but in cases where all the ID's cust_num = 'NONE' only then these should be marked as 'N'.
I've got a oracle install [non production, but devel] that is a tad screwed up. We moved the box and as a result changed the hostname to match the new naming scheme. Ever since then OracleEM has been somewhat confused. In anycase, I don't want OEM anyways now. Plan is to learn SQLplus.
That being said I've used emctl to shut down dbconsole, but it seems there is something somewhere that keeps restarting 2 processes that like to sit around and take up 100% cpu. I can kill them, they stay dead for a few hours then crop up again.I was able to find this out about them:
And then this, which caused me to conlucde its OracleEM:
SELECT sess.process, sess.status, sess.username, sess.schemaname, sql.sql_text FROM v$session sess, v$sql sql WHERE sql.sql_id(+) = sess.sql_id AND sess.process in (20334,20336) [code]...
i have three tables ot_cut_head,ot_cut_det and om_mc_master based on which fourth table ot_cut_opr and fifth table ot_cut_mc must get populated , Conditions are as follows
first one is based on job_no in ot_cut_head the selection criteria will be filtered,if the job number is like '%M' then type MISC will be chosen ,if job number is '%G' then GRAT TYPE will be picked from om_mc_master (Machine Master) and operations and machines based on this will be filtered.
Second all the cd_ps_desc will be taken from ot_cut_det and will be compared with om_mc_master to get their corresponding operation codes and machine codes , there can be 2 operations or 1 operation.
Finally if the match is found record will be inserted into ot_cut_opr and ot_cut_mc ,based on the criterias and what i want is the search criteria to be more flexible and if there are 2 operations 2 rows will be inserted and if one opeation is defined in om_mc_master ,then only one record will be inserted.
We have to make sure that if based on operation number stage will be populated ,if its first operation then stage will be 1 and if its second operation the stage will be 2.like previous operation also depends on them , the second operation will have the previous operation as first operation and so on.
CREATE TABLE om_mc_master ( mc_type VARCHAR2(12),mc_prof VARCHAR2(30),mc_prep_cd1 VARCHAR2(30),mc_mach_cd1 VARCHAR2 (30),mc_prep_cd2 VARCHAR2(30),mc_mach_cd2 VARCHAR2(30)); INSERT INTO OM_MC_MASTER VALUES ('MISC','TEE SCH','IR','HO','RE','HO'); insert into om_mc_master values('MISC','Vertical Brace','R','HM','I','HO'); insert into om_mc_master values('MISC','Pipe','IR','HO',NULL,NULL); INSERT INTO OM_MC_MASTER VALUES ('GRAT','PL','RE','HO',NULL,NULL); SQL> SELECT * FROM OM_MC_MASTER; [code]....
This is used the where clause of the REF CURSOR SELECT query which send the data back to SSRS
ie., SELECT BU.* FROM BU_DETAIL BU WHERE INSTR(V_BU_LST,BU_ID) <> 0;
INSTR has a chance to fail in this scenario if the value send from the front end is 123456,3456,4577
here 123456 does not exist in table, but it will be true for INSTR and values 1234 from table will be send back to SSRS which is wrong. Earlier I was using a function to convert the comma separated values to multi-rows and treat it like a lookup table.
But the main table has around million records , and each row has to processed against each row of lookup table, which makes it slower. To avoid this I used INSTR which is faster but can give wrong results.
I want to find the row with invalid day, month which are not matching with calendar day and month. Also the program should capture the data if the year <1900
with xx as (select 101 as ID, '24/05/1899' as create_date from dual union all select 101, '32/03/2012' from dual union all select 102 ,'30/02/2012' from dual union all select 101 , '29/02/2013' from dual
While starting up my database i am getting this error.by mistake i had moved log files, and after some time again i inserted those log files again into same directory i am getting this error.
UPDATE event_request SET REQ_EVENT_DATE = TO_CHAR(REQ_EVENT_DATE - INTERVAL '1' HOUR, 'yyyy-mm-dd hh24:mi') WHERE eventID=123
Oracle returns this error:
ERROR at line 4:ORA-01861: literal does not match format string
where line 4 is REQ_EVENT_DATE = TO_CHAR(REQ_EVENT_DATE - INTERVAL '1' HOUR, 'yyyy-mm-dd hh24:mi').The REQ_EVENT_DATE field was originally populated with this stripped down query:
