Sorted Target table

Hi all,
where/how can I use clause order by to get sorted data on target table?
I tried it via expression editor in mapping on some column in target datastore, but received something like "... the key word 'From' have not be found on the expected place..."
Forgive me this stupid question, I am just an old man new in ODI..
Thx

Hi,
Try with COL_NAME,
<%=snpRef.getColList("", " [COL_NAME]", ", ", "", "UD1") %>
FYR: Re: please suggest me
Thanks,
Guru

Similar Messages

How to sort columns in the target table

I have a simple mapping which I am trying to design. There's only one table on the source and one in the target . There are no filter conditions, only thing is I want the target table to be sorted.
Literally, say
Src is source table has 3 columns x,y,z
Trg is dest table and has 3 columns a,b,c
x--->a
y---->b
z---->c
The SQL should be
select x,y,z from src order by x,y.
I could do the mapping but the order by ..I could not do it .
IKM used: IKM BIAPPS Oracle Incremental Update

Why can't you use simple UPDATE command in EXECUTE SQL Task as below,
DROP TABLE SSN
DROP TABLE STAGING
DROP TABLE STUDENT
CREATE TABLE SSN(pn_id VARCHAR(100),ssn BIGINT)
INSERT INTO SSN VALUES('000616850',288258466)
INSERT INTO SSN VALUES('002160790',176268917)
CREATE TABLE Staging (ssn BIGINT, id INT, pn_id BIGINT, name VARCHAR(100), subject VARCHAR(100),grade VARCHAR(10), [academic year] INT, comments VARCHAR(100))
INSERT INTO Staging VALUES(288258466, 1001, '770616858','Sally Johnson', 'English','A', 2005,'great student')
INSERT INTO Staging VALUES(176268917, 1002, '192160792','Will Smith', 'Math','C', 2014,'no comments')
INSERT INTO Staging VALUES(444718562, 1003, '260518681','Mike Lira', 'Math','B', 2013,'no comments')
CREATE TABLE Student(id INT,pn_id BIGINT,subject VARCHAR(100), [academic year] INT, grade VARCHAR(10), comments VARCHAR(100) )
INSERT INTO Student VALUES(1001, '000616850', NULL,NULL,NULL ,NULL)
INSERT INTO Student VALUES(1002, '002160790', NULL,NULL,NULL ,NULL)
UPDATE Student SET Subject = C.Subject, [academic year]=C.[academic year], grade=C.grade,comments=C.comments
FROM SSN A INNER JOIN Student B
ON A.pn_id=B.pn_id INNER JOIN Staging C
ON A.ssn = C.ssn
SELECT * FROM Student
Regards, RSingh

Get Target table name and its count

Hi Experts,
I have created a Procedure where I want to get the Target table name and its count of the previous Interface in a Package and insert these values into the Audit table.
I am able to get the number of Inserts in previous step using getPrevStepLog() API, but I also want the Target table name and its total count of the previous Interface. Is there any generic way or code to bring these Information without hardcoding the table name. Kindly help to sort this issue.Thanks!
Warm Regards,
VBV

Hi VBV,
Please follow the below steps to audit ur execution.
Please note i coded in such a way that u can use ur existing procedure to capture the table name no need to query ur repository as its not advisable to query the repository as Oracle may change the structure in future which i personally feel right.
I created a Audit table with the below structure.
CREATE TABLE AUDIT_TABLE
INSERT_COUNT VARCHAR2(100 BYTE),
TABLE_NAME VARCHAR2(4000 BYTE)
Step 1 :
This step needs to be added in ur existing IKM after Commit step where i am using API to capture the target table name.
Step Name: Get Table Name
Command On Source
Technology: Oracle
Schema: Any Oracle related Schema
Command:
SELECT '<%=odiRef.getTargetTable("RES_NAME")%>' AS TGT_NAME FROM DUAL
Command On Target
Technology: Jython
Command:
TargetTable='#TGT_NAME'
Step 2:
In ur existing procedure add the below or add a new procedure.
Step Name: Audit Log
Command on Source
Technology : Oracle
Schema: Whichever schema holds the audit table
Command on Target
Technology : Jython
Command:
import java.sql as sql
import java.lang as lang
myCon =odiRef.getJDBCConnection("SRC")
MyStmt=myCon.createStatement()
myRs = MyStmt.executeQuery("INSERT INTO ODITGT.AUDIT_TABLE (INSERT_COUNT,TABLE_NAME) VALUES (<%=odiRef.getPrevStepLog("INSERT_COUNT")%>,'"+TargetTable+"')")
Thats it.
OTN doesnt allow me to put + sign in front of any word ( its assume its a Italic) so please add + (plus) sign before and after TargetTable in the above script.
This way u can capture the table name, insert count in to ur audit table.
Please note u need to call this Procedure after the successful completion of the interface in a package and make sure u selected the right IKM which capture the table name as in Step 1.
I tested in my local and its working fine. Let me know if u find any difficulties in implementation.
Thanks,
Guru

Is it possible to run an update via target tables post load command

Is it possible to run a sql update statement via post load command of target table?
What does it mean by post load command, does it gets triggered only after a successful execution of a data flow (to Target table)?
Or it can also gets triggered even if a dataflow fails?
this is teh sql which i would like to run via post if dataflow success only.
Update tbl_job_status set end_time=nvl(NULL,sysdate()) where job_id = (select max job_id from tbl_job_status where j_dataflowname='Chk_Order_Delta')

I do not believe it will necessary run if the df fails. If you want your job status table updated no matter what (apart from a complete DS meltdown), you'll want to wrap things in a try-catch to make sure you're handling the update. You can do it in a script, or with a dataflow. A nice way to handle all this sort of operational metadata tracking stuff is to write yourself a library of custom functions and then consistently call them fore & aft of jobs or workflows or whatever components you want to track. If you've got access to a Rapid Mart, you might look at what they've done there -- nice stuff (although you can't just "borrow" it).

Can not Load CLOB data 32k to target table

SQL> DESC testmon1 ;
Name Null? Type
FILENAME VARCHAR2(200)
SCANSTARTTIME VARCHAR2(50)
SCANENDTIME VARCHAR2(50)
JOBID VARCHAR2(50)
SCANNAME VARCHAR2(200)
SCANTYPE VARCHAR2(200)
FAULTLINEID VARCHAR2(50)
RISK VARCHAR2(5)
VULNNAME VARCHAR2(2000)
CVE VARCHAR2(200)
DESCRIPTION CLOB
OBSERVATION CLOB
RECOMMENDATION CLOB
SQL> DESC test_target;
Name Null? Type
LOCALID NOT NULL NUMBER
DESCRIPTION NOT NULL CLOB
SCANTYPE NOT NULL VARCHAR2(12)
RISK NOT NULL VARCHAR2(6)
TIMESTAMP NOT NULL DATE
VULNERABILITY_NAME NOT NULL VARCHAR2(2000)
CVE_ID VARCHAR2(200)
BUGTRAQ_ID VARCHAR2(200)
ORIGINAL VARCHAR2(50)
RECOMMEND CLOB
VERSION VARCHAR2(15)
FAMILY VARCHAR2(15)
XREF VARCHAR2(15)
create or replace PROCEDURE proc1 AS
CURSOR C1 IS
SELECT FAULTLINEID,VULNNAME,scanstarttime, risk,
dbms_lob.substr(DESCRIPTION,dbms_lob.getlength(DESCRIPTION),1) "DESCR",dbms_lob.substr(OBSERVATION,dbms_lob.getlength(OBSERVATION),1) "OBS",
dbms_lob.substr(RECOMMENDATION) "REC",CVE
FROM testmon1;
c_rec C1%ROWTYPE;
descobs clob;
FSCAN_VULN_TRANS_REC VULN_TRANSFORM_STG%ROWTYPE;
TIMESTAMP varchar2(50);
riskval varchar2(10);
pCTX PLOG.LOG_CTX := PLOG.init (pSECTION => 'foundscanVuln Procedure',
pLEVEL => PLOG.LDEBUG,
pLOG4J => TRUE,
pLOGTABLE => TRUE,
pOUT_TRANS => TRUE,
pALERT => TRUE,
pTRACE => TRUE,
pDBMS_OUTPUT => TRUE);
amount number;
buffer varchar2(32000);
BEGIN
---INITIALIZE THE LOCATOR FOR CLOB DATA TYPE
select observation into descobs from testmon1 where rownum=1;
OPEN C1;
loop
fetch C1 INTO c_rec;
exit when C1%NOTFOUND;
--LOAD THE DESCRIPTION FIELD FROM CURSOR AND WRITE IT TO THE CLOB LOCATOR descobs.
dbms_lob.Write(descobs,dbms_lob(c_rec.DESCR),1,c_rec.DESCR);
------APPEND THE OBSERVATION FIELD FROM CURSOR TO THE CLOB LOCATOR descobs.
dbms_lob.Writeappend(descobs,dbms_lob(c_rec.DESCR),c_rec.OBS);
-- dbms_output.put_line ('the timestamp is :'||c_rec.scanstarttime);
--dbms_lob.write(descobs,amount,1,buffer);
descobs:=c_rec.OBS;
--dbms_lob.read(descobs,amount,1,buffer);
--dbms_lob.append(c_rec.DESCR,c_rec.OBS);
--descobs:=c_rec.OBS;
--dbms_output.put_line ('the ADDED DESCROBS is :'||dbms_lob.substr(c_rec.DESCR,dbms_lob.getlength(c_rec.DESCR),1));
dbms_output.put_line ('the ADDED DESCRIPTION AND OBSERVATION is :'||descobs);
--dbms_output.put_line ('the DESCROBS buffer is :'||buffer);
SELECT DESCRIPTION INTO FSCAN_VULN_TRANS_REC.DESCRIPTION
FROM TESTMON1 WHERE ROWNUM=1;
---------LOAD THE DESCRIPTION+ observation value into the target table description
DBMS_LOB.WRITE(FSCAN_VULN_TRANS_REC.DESCRIPTION, dbms_lob.getlength(descobs),1,descobs);
TIMESTAMP:=substr(c_rec.scanstarttime,1,10)||' '|| substr(c_rec.scanstarttime,12,8);
IF c_rec.risk <3
THEN riskval:='Low';
ELSIF c_rec.risk <6
THEN riskval:='Medium';
ELSIF c_rec.risk <10
THEN riskval:='High';
END IF;
FSCAN_VULN_TRANS_REC.TIMESTAMP:=TO_DATE(TIMESTAMP, 'YYYY/MM/DD HH24:MI:SS');
FSCAN_VULN_TRANS_REC.risk:= riskval;
--dbms_lob.append(c_rec.DESCR,c_rec.OBS);
FSCAN_VULN_TRANS_REC.DESCRIPTION:=c_rec.DESCR;
FSCAN_VULN_TRANS_REC.RECOMMEND:=c_rec.REC;
FSCAN_VULN_TRANS_REC.LocalID:=to_number(c_rec.FAULTLINEID);
FSCAN_VULN_TRANS_REC.SCANTYPE:='FOUNDSCAN';
FSCAN_VULN_TRANS_REC.CVE_ID:=c_rec.CVE;
FSCAN_VULN_TRANS_REC.VULNERABILITY_NAME:=c_rec.VULNNAME;
-- dbms_output.put_line ('the plog timestamp is :'||timestamp);
-- dbms_output.put_line ('the timestamp is :'||riskval);
--dbms_output.put_line ('the recommend is :'||FSCAN_VULN_TRANS_REC.RECOMMEND);
--dbms_output.put_line ('the app desc is :'||FSCAN_VULN_TRANS_REC.DESCRIPTION);
insert into test_target values FSCAN_VULN_TRANS_REC;
End loop;
close C1;
commit;
EXCEPTION
WHEN OTHERS THEN
-- dbms_output.put_line ('Data not found');
-----------dbms_output.put_line (sqlcode|| ':'||sqlerrm);
end proc1;
using dbms_lob package is not helping. Either DB stops responding. Or the Observation field ( which has max length >300000) can not be loaed into a CLOB variable.
Please help or give me a sample code that helps.

select
     BANKING_INSTITUTION.BANK_REF_CODE     C1_BANK_ID,
     BANKING_INSTITUTION.NAME_BANK     C2_BANK_NAME,
     BANKING_INSTITUTION.BANK_NUMBER     C3_BANK_NUMBER,
     BANKING_INSTITUTION.ISO_CODE     C4_GBA_CODE,
     BANKING_INSTITUTION.STATUS     C5_STATUS,
     BANKING_INSTITUTION.SOURCE     C6_SOURCE,
     BANKING_INSTITUTION.START_DATE_BANK     C7_START_DATE,
     BANKING_INSTITUTION.ADDRESS_BANK     C8_BANK_ADDRESS1
from     REF_DATA_DB.BANKING_INSTITUTION BANKING_INSTITUTION
where     (1=1)
insert /*+ append */ into XXSVB.C$_0XXSVB_BANKS_STAGING
     C1_BANK_ID,
     C2_BANK_NAME,
     C3_BANK_NUMBER,
     C4_GBA_CODE,
     C5_STATUS,
     C6_SOURCE,
     C7_START_DATE,
     C8_BANK_ADDRESS1
values
     :C1_BANK_ID,
     :C2_BANK_NAME,
     :C3_BANK_NUMBER,
     :C4_GBA_CODE,
     :C5_STATUS,
     :C6_SOURCE,
     :C7_START_DATE,
     :C8_BANK_ADDRESS1
)

Null value in target table

Hi All
I am using ODI 10.1.13
I have declred teh Target datastore
and mapped the target datastore with source in teh interface
But when ODI is creating teh target database all the columns are created with accepting NULL Properties.
where as i hv declared some as null and some with notnull.
when i checked in IKM Oracle Incrimental KM(MERGE)-> create target table step
the coding is
create table <%=odiRef.getTable("L", "TARG_NAME", "A")%>
<%=odiRef.getTargetColList("", "[COL_NAME]\t[DEST_CRE_DT] NULL", ",\n\t", "")%>
due to which the target table is created with all field havine NULL properties.
But i want the columns should hv the properties declared in the datastore.Eg. some columns with NULL and some columns with NOT NULL properties.
Please help..
Gourisankar

In the IKM Merge , change the null to + odiRef.getInfo("DEST_DDL_NULL") so finally something like this <%=odiRef.getTargetColList("", "[COL_NAME]\t[DEST_CRE_DT] " + odiRef.getInfo("DEST_DDL_NULL"), ",\n\t", "")%> For bulk insert try SQL Control Append

Table compare deleting rows which does not exist in target table

Hi Gurus,
I am struggling with an issue in Data Services.
I have a job which uses Table Compare, then History Preserving and then a Key Generation transforms.
There is every possibility that data would get deleted from the source table.
Now, I want to delete them from the target table also.
I tried Detect deleted rows but it is not working.
Could some one please help me on this issue.
Thanks,
Raviteja.

Doesn't history preserving really only operate on "Update" rows. Wouldn't it only process the deletes if you turned the "Preserve Delete row(s) as update row(s)" on?
I would think if you turned on Detect Delete rows in the Table compare and did not turn this on in the history preserving it would retain those rows as delete rows and effectively remove them from the target.
Preserve delete row(s) as update row(s)
Converts DELETE rows to UPDATE rows in the target warehouse and, if you previously set effective date values (Valid from and Valid to), sets the Valid To value to the execution date. Use this option to maintain slowly changing dimensions by feeding a complete data set first through the Table Comparison transform with its Detect deleted row(s) from comparison table
option selected.

How to load data into 3 different target tables usin BODS ?

Hello Friends,
I have 5 different source tables with same field definitions, Now I want to load all the records into three/four different target tables (Flat file, SQL Server, XML, and Oracle ) Could anyone please tell me how to do this task ?
Thanks in Advance,
Bheem.

Hello Bheem,
You can create separated dataflow for each target as suggested by Bala, this is a good choice when evaluating you scenario.
If you put all targets in the same dataflow you may experience problems if one of them is down (as you have different servers as targets).
BODS will send the data simultaneously to all targets and when one fails, the load will be break and you may have the tables not sinc anyways (if you plan to use a single dataflow to avoid this situation, it won't do it).
So the best option is to create separated dataflow and put inside another one so you can run a singe dataflow that will call each one of the target and if onw fails the other will complete accordingly.
Then if you put them inside a error trapping, you may even make your dataflow to retry the load prior abend the job.
Think about cascading your dataflows and let the leaf level simple as it can be so it will be easier to schedule and debug your process.
Pay attention to datatypes and other conversions you might need when working with more than one source/target.
Regards,

How to delete some date in target table at a mapping?

How to delete some date in target table at a mapping?
I extract date from source tabel into target table,
but before extract date I want to delete some date from target?
how to do?

Just to change a bit of terminology in the reply, within the mapping, click on operator properties and choose TRUNCATE/INSERT.
Note that truncate is dependent on constraints, so you probably must disable those before doing this. You can of course do DELETE/INSERT...
Jean-Pierre

How do I create a target table with the same PK as the source table?

I am trying to create a target table in a mapping that will end up with the same primary key as the source table.
It is a simple map that simply uses a subset of the columns of the source table in the target table. I was wanting to create and bind a new table by dragging the columns I want from the source to the initially blank target table operator, change the column names and create a primary key to match the source table.
I can't seem to be able to create a constraint on the table in the mapping. I can create the constraint after the table is created and boound to the database object but the PK doesn't carry back into the mapping.
I need it in the mapping so I can use the UPDATE/INSERT operation and use the 'All Constraints' implementation. The mapping won't let me validate the object without the PK on it in the map.
Believe it or not folks, I am getting better at this.
Thanks very much for the guidance.
Gary

Hi Gary
You are close, you are really close... :-))
You need to do exactly as you propose plus one extra step. Build the map as you describe, binding the new table to the target. Then you edit the table definition to add the primary key and any other constraints you need. After this is the step that you are missing.
You need to do the following:
1. Go back and re-edit the map
2. Right click on the table
3. From the pop up menu, select Reconcile Inbound
4. Set any operators that you need for the UPDATE/INSERT
5. Save the map
6. Commit your changes
The first three steps above make the map read in the indexes and constraints that you set on the table. Finally, you need to deploy the table and then deploy the map.
Hope this helps
Regards
Michael

How to get source table name according to target table

hi all
another question:
once a map was created and deployed,the corresponding information was stored in the repository and rtr repository.My question is how to find the source table name according to the target table,and in which table these records are recorded.
somebody help me plz!!
thanks a lot!

This is a query that will get you the operators in a mapping. To get source and targets you will need some additional information but this should get you started:
set pages 999
col PROJECT format a20
col MODULE format a20
col MAPPING format a25
col OPERATOR format a20
col OP_TYPE format a15
select mod.project_name PROJECT
, map.information_system_name MODULE
, map.map_name MAPPING
, cmp.map_component_name OPERATOR
, cmp.operator_type OP_TYPE
from all_iv_xform_maps map
, all_iv_modules mod
, all_iv_xform_map_components cmp
where mod.information_system_id = map.information_system_id
and map.map_id = cmp.map_id
and mod.project_name = '&Project'
order by 1,2,3
Jean-Pierre

How to Assign a default value and Increement in my Target table ........?

Hi,
In my target table I want to Increment a value and store it in the table ........
the increement shud start frm ...........5001,5002,5003..................etc
How can I achieve this...........?
I created an numeric Variable and in assigned the default as 5000 nad action as Historize.........
In the target Datastore ......I assigned VAR+1.......
but in the Target column I got 5001 for all the columns........
Which way is the best way to deal with the Variables.......
Thanks
AK
Message was edited by:
AK2008

Hi Gurusank and Micropole,
Thanks for reply and info.....
I am trying out one of the solutions first........
Gurusank,
I have created a sequence........whether I need to mention the schema and table name ,column.........also ( I gave all the information) and Increement column =1 and saved it .......
Then I follwed the procedure as u have mentioned .....
And in the Target field also I gave the same the same.......for tat column I gave choose (target for execution).
And I directly ran the interface ..........
I got the folowing Error:
287 : 42000 : java.sql.SQLException: ORA-02287: sequence number not allowed here
java.sql.SQLException: ORA-02287: sequence number not allowed here
     at oracle.jdbc.driver.DatabaseError.throwSqlException(DatabaseError.java:125)
     at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:316)
     at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:282)
     at oracle.jdbc.driver.T4C8Oall.receive(T4C8Oall.java:639)
     at oracle.jdbc.driver.T4CPreparedStatement.doOall8(T4CPreparedStatement.java:185)
     at oracle.jdbc.driver.T4CPreparedStatement.execute_for_rows(T4CPreparedStatement.java:633)
     at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1086)
     at oracle.jdbc.driver.OraclePreparedStatement.executeInternal(OraclePreparedStatement.java:2984)
     at oracle.jdbc.driver.OraclePreparedStatement.executeUpdate(OraclePreparedStatement.java:3057)
     at com.sunopsis.sql.SnpsQuery.executeUpdate(SnpsQuery.java)
     at com.sunopsis.dwg.dbobj.SnpSessTaskSql.execStdOrders(SnpSessTaskSql.java)
     at com.sunopsis.dwg.dbobj.SnpSessTaskSql.treatTaskTrt(SnpSessTaskSql.java)
     at com.sunopsis.dwg.dbobj.SnpSessTaskSqlI.treatTaskTrt(SnpSessTaskSqlI.java)
     at com.sunopsis.dwg.dbobj.SnpSessTaskSql.treatTask(SnpSessTaskSql.java)
     at com.sunopsis.dwg.dbobj.SnpSessStep.treatSessStep(SnpSessStep.java)
     at com.sunopsis.dwg.dbobj.SnpSession.treatSession(SnpSession.java)
     at com.sunopsis.dwg.cmd.DwgCommandSession.treatCommand(DwgCommandSession.java)
     at com.sunopsis.dwg.cmd.DwgCommandBase.execute(DwgCommandBase.java)
     at com.sunopsis.dwg.cmd.e.i(e.java)
     at com.sunopsis.dwg.cmd.g.y(g.java)
     at com.sunopsis.dwg.cmd.e.run(e.java)
     at java.lang.Thread.run(Thread.java:534)
Could u plz let me know where I am doing the mistake
Thanks
AK

Insert/select one million rows at a time from source to target table

Hi,
Oracle 10.2.0.4.0
I am trying to insert around 10 million rows into table target from source as follows:
INSERT /*+ APPEND NOLOGGING */ INTO target
SELECT *
FROM source f
WHERE
        NOT EXISTS(SELECT 1 from target m WHERE f.col1 = m.col2 and f.col2 = m.col2);There is a unique index on target table on col1,col2
I was having issues with undo and now I am getting the follwing error with temp space
ORA-01652: unable to extend temp segment by 64 in tablespace TEMPI believce it would be easier if I did bulk insert one million rows at a time and commit.
I appriciate any advice on this please.
Thanks,
Ashok

902986 wrote:
NOT EXISTS(SELECT 1 from target m WHERE f.col1 = m.col2 and f.col2 = m.col2);
I don't know if it has any bearing on the case, but is that WHERE clause on purpose or a typo? Should it be:
        NOT EXISTS(SELECT 1 from target m WHERE f.col1 = m.COL1 and f.col2 = m.col2);Anyway - how much of your data already exists in target compared to source?
Do you have 10 million in source and very few in target, so most of source will be inserted into target?
Or do you have 9 million already in target, so most of source will be filtered away and only few records inserted?
And what is the explain plan for your statement?
INSERT /*+ APPEND NOLOGGING */ INTO target
SELECT *
FROM source f
WHERE
        NOT EXISTS(SELECT 1 from target m WHERE f.col1 = m.col2 and f.col2 = m.col2);As your error has to do with TEMP, your statement might possibly try to do a lot of work in temp to materialize the resultset or parts of it to maybe use in a hash join before inserting.
So perhaps you can work towards an explain plan that allows the database to do the inserts "along the way" rather than calculate the whole thing in temp first.
That probably will go much slower (for example using nested loops for each row to check the exists), but that's a tradeoff - if you can't have sufficient TEMP then you may have to optimize for less usage of that resource at the expense of another resource ;-)
Alternatively ask your DBA to allocate more room in TEMP tablespace. Or have the DBA check if there are other sessions using a lot of TEMP in which case maybe you just have to make sure your session is the only one using lots of TEMP at the time you execute.

Don't have primary Key in Target table getting errror while creating index

Hi All,
I don't have primary key column in target table, while exicuting mapping I got a error while creating INDEX.
Could you please help how to slove this

Hi,
That is a process definition issue.
If you don't have a PK then:
1) or you don't execute updates
2)or you have an alternate Key to update it.
Case 1) just change the KM to IKM Control Append
Case 2) at interface, go to each column what is Alternate Key and check it as key (click at column and check the box Key at bottom of propriety window).
Does it work to you?

A question about transaction consistency between multible target tables

Dear community,
My replication is ORACLE 11.2.3.7->ORACLE 11.2.3.7 both running on linux x64 and GG version is 11.2.3.0.
I'm recovering from an error when trail file was moved away while dpump was writing to it.
After moving the file back dpump abended with an error
2013-12-17 11:45:06 ERROR   OGG-01031 There is a problem in network communication, a remote file problem, encryption keys for target and source do
not match (if using ENCRYPT) or an unknown error. (Reply received is Expected 4 bytes, but got 0 bytes, in trail /u01/app/ggate/dirdat/RI002496, seqno 2496,
reading record trailer token at RBA 12999993).
I googled for It and found no suitable solution except for to try "alter extract <dpump>, entrollover".
After rolling over trail file replicat as expected ended with
REPLICAT START 1
2013-12-17 17:56:03 WARNING OGG-01519 Waiting at EOF on input trail file /u01/app/ggate/dirdat/RI002496, which is not marked as complete;
but succeeding trail file /u01/app/ggate/dirdat/RI002497 exists. If ALTER ETROLLOVER has been performed on source extract,
ALTER EXTSEQNO must be performed on each corresponding downstream reader.
So I've issued "alter replicat <repname>, extseqno 2497, extrba 0" but got the following error:
REPLICAT START 2
2013-12-17 18:02:48 WARNING OGG-00869 Aborting BATCHSQL transaction. Detected inconsistent result:
executed 50 operations in batch, resulting in 47 affected rows.
2013-12-17 18:02:48 WARNING OGG-01137 BATCHSQL suspended, continuing in normal mode.
2013-12-17 18:02:48 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
2013-12-17 18:02:48 WARNING OGG-01004 Aborted grouped transaction on 'M.CLIENT_REG', Database error
1403 (OCI Error ORA-01403: no data found, SQL <UPDATE "M"."CLIENT_REG" SET "CLIENT_CODE" =
:a1,"CORE_CODE" = :a2,"CP_CODE" = :a3,"IS_LOCKED" = :a4,"BUY_SUMMA" = :a5,"BUY_CHECK_CNT" =
:a6,"BUY_CHECK_LIST_CNT" = :a7,"BUY_LAST_DATE" = :a8 WHERE "CODE" = :b0>).
2013-12-17 18:02:48 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
2013-12-17 18:02:48 WARNING OGG-01154 SQL error 1 mapping LS.CHECK to M.CHECK OCI Error ORA-00001:
unique constraint (M.CHECK_PK) violated (status = 1). INSERT INTO "M"."CHECK"
("CODE","STATE","IDENT_TYPE","IDENT","CLIENT_REG_CODE","SHOP","BOX","NUM","KIND","KIND_ORDER","DAT","SUMMA","LIST_COUNT","RETURN_SELL_CHECK_CODE","RETURN_SELL_SHOP","RETURN_SELL_BOX","RETURN_SELL_NUM","RETURN_SELL_KIND","INSERTED","UPDATED","REMARKS")
VALUES
(:a0,:a1,:a2,:a3,:a4,:a5,:a6,:a7,:a8,:a9,:a10,:a11,:a12,:a13,:a14,:a15,:a16,:a17,:a18,:a19,:a20).
2013-12-17 18:02:48 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
The report stated the following:
Reading /u01/app/ggate/dirdat/RI002497, current RBA 1149, 0 records
Report at 2013-12-17 18:02:48 (activity since 2013-12-17 18:02:46)
From Table LS.MK_CHECK to LSGG.MK_CHECK:
       #                   inserts:         0
       #                   updates:         0
       #                   deletes:         0
       #                  discards:         1
From Table LS.MK_CHECK to LSGG.TL_MK_CHECK:
       #                   inserts:         0
       #                   updates:         0
       #                   deletes:         0
       #                  discards:         1
At that time I came to the conclusion that using etrollover was not a good idea Nevertheless I had to upload my data to perform consistency check.
My mapping templates are set up as the following:
LS.CHECK->M.CHECK
LS.CHECK->M.TL_CHECK
(such mapping is set up for every table that is replicated).
TL_CHECK is a transaction log, as I name it,
and this peculiar mapping is as the following:
ignoreupdatebefores
map LS.CHECK, target M.CHECK, nohandlecollisions;
ignoreupdatebefores
map LS.CHECK, target M.TL_CHECK ,colmap(USEDEFAULTS,
FILESEQNO = @GETENV ("RECORD", "FILESEQNO"),
FILERBA = @GETENV ("RECORD", "FILERBA"),
COMMIT_TS = @GETENV( "GGHEADER", "COMMITTIMESTAMP" ),
FILEOP = @GETENV ("GGHEADER","OPTYPE"), CSCN = @TOKEN("TKN-CSN"),
RSID = @TOKEN("TKN-RSN"),
OLD_CODE = before.CODE
, OLD_STATE = before.STATE
, OLD_IDENT_TYPE = before.IDENT_TYPE
, OLD_IDENT = before.IDENT
, OLD_CLIENT_REG_CODE = before.CLIENT_REG_CODE
, OLD_SHOP = before.SHOP
, OLD_BOX = before.BOX
, OLD_NUM = before.NUM
, OLD_NUM_VIRT = before.NUM_VIRT
, OLD_KIND = before.KIND
, OLD_KIND_ORDER = before.KIND_ORDER
, OLD_DAT = before.DAT
, OLD_SUMMA = before.SUMMA
, OLD_LIST_COUNT = before.LIST_COUNT
, OLD_RETURN_SELL_CHECK_CODE = before.RETURN_SELL_CHECK_CODE
, OLD_RETURN_SELL_SHOP = before.RETURN_SELL_SHOP
, OLD_RETURN_SELL_BOX = before.RETURN_SELL_BOX
, OLD_RETURN_SELL_NUM = before.RETURN_SELL_NUM
, OLD_RETURN_SELL_KIND = before.RETURN_SELL_KIND
, OLD_INSERTED = before.INSERTED
, OLD_UPDATED = before.UPDATED
, OLD_REMARKS = before.REMARKS), nohandlecollisions, insertallrecords;
As PK violation fired for CHECK, I've changed nohandlecollisions to handlecollisions for LS.CHECK->M.CHECK mapping and restarted an replicat.
To my surprise it ended with the following error:
REPLICAT START 3
2013-12-17 18:05:55 WARNING OGG-00869 Aborting BATCHSQL transaction. Database error 1 (ORA-00001:
unique constraint (M.CHECK_PK) violated).
2013-12-17 18:05:55 WARNING OGG-01137 BATCHSQL suspended, continuing in normal mode.
2013-12-17 18:05:55 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
2013-12-17 18:05:55 WARNING OGG-00869 OCI Error ORA-00001: unique constraint (M.PK_TL_CHECK)
violated (status = 1). INSERT INTO "M"."TL_CHECK"
("FILESEQNO","FILERBA","FILEOP","COMMIT_TS","CSCN","RSID","CODE","STATE","IDENT_TYPE","IDENT","CLIENT_REG_CODE","SHOP","BOX","NUM","KIND","KIND_ORDER","DAT","SUMMA","LIST_COUNT","RETURN_SELL_CHECK_CODE","RETURN_SELL_SHOP","RETURN_SELL_BOX","RETURN_SELL_NUM","RETURN_SELL_KIND","INSERTED","UPDATED","REMARKS")
VALUES
(:a0,:a1,:a2,:a3,:a4,:a5,:a6,:a7,:a8,:a9,:a10,:a11,:a12,:a13,:a14,:a15,:a16,:a17,:a18,:a19,:a20,:a21,:a22,:a23,:a24,:a25,:a26).
2013-12-17 18:05:55 WARNING OGG-01004 Aborted grouped transaction on 'M.TL_CHECK', Database error 1
(OCI Error ORA-00001: unique constraint (M.PK_TL_CHECK) violated (status = 1). INSERT INTO
"M"."TL_CHECK"
("FILESEQNO","FILERBA","FILEOP","COMMIT_TS","CSCN","RSID","CODE","STATE","IDENT_TYPE","IDENT","CLIENT_REG_CODE","SHOP","BOX","NUM","KIND","KIND_ORDER","DAT","SUMMA","LIST_COUNT","RETURN_SELL_CHECK_CODE","RETURN_SELL_SHOP","RETURN_SELL_BOX","RETURN_SELL_NUM","RETURN_SELL_KIND","INSERTED","UPDATED","REMARKS")
VALUES
(:a0,:a1,:a2,:a3,:a4,:a5,:a6,:a7,:a8,:a9,:a10,:a11,:a12,:a13,:a14,:a15,:a16,:a17,:a18,:a19,:a20,:a21,:a22,:a23,:a24,:a25,:a26)).
2013-12-17 18:05:55 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
2013-12-17 18:05:55 WARNING OGG-01154 SQL error 1 mapping LS.CHECK to M.TL_CHECK OCI Error
ORA-00001: unique constraint (M.PK_TL_CHECK) violated (status = 1). INSERT INTO "M"."TL_CHECK"
("FILESEQNO","FILERBA","FILEOP","COMMIT_TS","CSCN","RSID","CODE","STATE","IDENT_TYPE","IDENT","CLIENT_REG_CODE","SHOP","BOX","NUM","KIND","KIND_ORDER","DAT","SUMMA","LIST_COUNT","RETURN_SELL_CHECK_CODE","RETURN_SELL_SHOP","RETURN_SELL_BOX","RETURN_SELL_NUM","RETURN_SELL_KIND","INSERTED","UPDATED","REMARKS")
VALUES
(:a0,:a1,:a2,:a3,:a4,:a5,:a6,:a7,:a8,:a9,:a10,:a11,:a12,:a13,:a14,:a15,:a16,:a17,:a18,:a19,:a20,:a21,:a22,:a23,:a24,:a25,:a26).
2013-12-17 18:05:55 WARNING OGG-01003 Repositioning to rba 1149 in seqno 2497.
I've expected that batchsql will fail cause it does not support handlecollisions, but I really don't understand why any record was inserted into TL_CHECK and caused PK violation, cause I thought that GG guarantees transactional consistency and that any transaction that caused an error in _ANY_ of target tables will be rollbacked for _EVERY_ target table.
TL_CHECK has PK set to (FILESEQNO, FILERBA), plus I have a special column that captures replication run number and it clearly states that a record causing PK violation was inserted during previous run (REPLICAT START 2).
BTW report for the last shows
Reading /u01/app/ggate/dirdat/RI002497, current RBA 1149, 1 records
Report at 2013-12-17 18:05:55 (activity since 2013-12-17 18:05:54)
From Table LS.MK_CHECK to LSGG.MK_CHECK:
       #                   inserts:         0
       #                   updates:         0
       #                   deletes:         0
       #                  discards:         1
From Table LS.MK_CHECK to LSGG.TL_MK_CHECK:
       #                   inserts:         0
       #                   updates:         0
       #                   deletes:         0
       #                  discards:         1
So somebody explain, how could that happen?

Write the query of the existing table in the form of a function with PRAGMA AUTONOMOUS_TRANSACTION.
examples here:
http://www.morganslilbrary.org/library.html

Sorted Target table

Similar Messages

Maybe you are looking for