Help in data synchonization

Question

Hi All,
I have a two databases with same table structure. DB - A and DB - B. Table is WEB_DESPATCHES
DB - B is where transaction takes place and i move that data to DB - A every day through a procedure from DB - B.
The query I user is
select <columns> from WEB_DESPATCHES@DB - B
minus
select <columns> from WEB_DESPATCHES@DB - A
so that differential data gets inserted into DB - A.
Now my question is suppose some columns gets updated in DB - B for previous data so that how should i track and update only those columns data in DB - A
as for these records i have already taken data but data in DB - A and DB - B mismatches?
Say column SHPD_QTY,PKMS_CARTON_NBR,UNITS-PKD gets updated later in future ( say 10 days later)
How should I do in such case, Please suggest?
Table struct.
SQL> desc web_despatches
Name            Type         Nullable Default Comments
PKT_CTRL_NBR    VARCHAR2(20) Y                        
ORDER_NUMBER    VARCHAR2(30) Y                        
ORDER_REFERENCE VARCHAR2(30) Y                        
ORDER_LINE      VARCHAR2(4)  Y                        
ITEM_CODE       VARCHAR2(15) Y                        
SHPD_QTY        NUMBER(7)    Y                        
PKT_QTY         NUMBER(7)    Y                        
CONSIGNMENT_NO  VARCHAR2(20) Y                        
COPOS_LINE_NO   VARCHAR2(10) Y                        
COPOS_ID        VARCHAR2(10) Y                        
QTY_ORDERED     NUMBER(7)    Y                        
SKU_ORDERED     VARCHAR2(10) Y                        
DATE_CREATED    DATE         Y                        
STATUS          NUMBER       Y                        
PKMS_CARTON_NBR VARCHAR2(20) Y                        
UNITS_PAKD      NUMBER       Y                        

Answer 1

Sid_ Z. wrote:
My question is
Currently I am using this query
insert into web_despatches
select * from web_despatches@nrstp
minus
select * from web_despatches;
So after populating data into web_despatches, if suppose data is updated(few columns) in web_despatches@nrstp
then how should i keep track of this data as above statement doesn't return any rows except new ones.That is your solution to the problem and not the problem itself. Why is the web dispatches table split across 2 databases? What are the sizes? Are one the master and the other the duplicate? Are they peers? Why does rows from one table need to be pulled into the other?
There are numerous way to address problems in Oracle. Materialised views are one option that comes to mind that can address such a problem. Change Data Capture is another. Message Queues. Etc.
Your solution... I dislike. The reason is that this requires a lot of network traffic between the two servers to perform that SQL. This does not scale. The bigger the 2 tables. The more traffic. The slower the response.
Networks by nature tends to get slower and slower. Not faster. Same network infrastructure, but growing number of servers and client, growing amount of web traffic, e-mail traffic, database traffic and so on. So even if your data volumes stay constant, your SQL performance will still decrease over time as the network is impacted by other applications and systems.
So if data is to be distributed, it has to be done with a cunning plan.. (and perhaps with a giant turnip or two... {noformat};-){noformat} ). Unfortunately your SQL is not cunning, not even by the standards of good o'l Baldrick.