on 10-13-2008 6:55 PM
Hello All,
I have one big question regarding the INIT load. Let me explain you my question with an example:
I have a datasource say 2LIS_02_ITM for which I ran the setup tables on JAN 2008. The setup table is filled with say 10,000 records. After that we are running Delta loads everyday till today (almost 30,000 records). Everything is going fine.
Imagine I lost a delta (delta failed and did'nt I do a repeat delta for 3 days, by then next delta came), Hence I am planning to do a Init load again without deleting / filling the setup tables. Now my question is :
1) Will I get the 10,000 records which got filled initially in the setup tables or will I get 40,000 records (that 10,000 + 30,000 records)
2) If it bring 40,000 how is that possible as I hav'nt filled the setup tables?
3) Is it each time I need to fill the setup tables to start the Init load.
Waiting for your guidance.........
Hello
I do not think you have to do an init. How many delta's have come in since the last Delta failed?
My guess is the delta's that came in after that failed delta must be repeats - pls check.
What you can do is to correct the last request that came in (assuming it was also a repeat). And if it was a repeat it must have brought in its own records + records of last delta.
To correct the last repeat this is what you can do:
Mark QM status of that request red. Delete the requests and all requests after that request from Data targets.
Do a repeat of this delta request.
I think you should be fine after that. A delta can be able to get back your data. There is no need for init unless this 3 day gap has resulted in huge loss of data.
Hope this helps.
Edited by: BI Quest on Oct 13, 2008 8:42 PM
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello BI Quest,
Almost three delta's (with records) came after that failed delta (Friday). This happened on the week end. Even if you make the status of the last delta red , it will only repeat that delta and it wont pull the Friday's delta as I did'nt do the repeat delta for the failed friday delta.
Also can you please answer my above question :
Do we need to fill the set up tables each time we run a INIT load ??????
Yes...to answer your question
But...agin I will suggest not to go by that method. Only one delta had failed so why do you want to wipe away the timestamp of the last init just becasue of 1 delta failure? it is possible to correct it
First of all i hope you have stopped the process chain or put a hold to Delta jobs
Now, To answer your doubt, about the status red ..See you mark the status red in each of the the data target before delting the failed request from there (I usually do not just delete the requests frm data targets - I mark them red and then delte it to be safe).
Do not forget to delete subsequent requests afterthe failed delta (agin I always mark them red before doing so)
Regarding the actual reuest itself...you need not mark it red in the monitor..You have got to correct this request and successful - it will automatically turn green.
If after successfully correcting the request, it still does not turn green, you can set qm status green. The next delta will recognize this a successful request and bring in the next delta (this happens in cases where you have the infopackage in a process chian when it shows green because subsequent processes were not completed etc).
Let me know in case you have any questions.
Hi Ram,
Repair Full request is similar to Full load, we can load data based on selections.
Repair full request will not distrub the existing delta mechanism,
where ad after running delta loads, if we run Full load it will impact Delta loads, next delta load will fail after full load.
Regards
Daya Sagar
Hi Ram,
We use full repair (instead of just FUll load) when there is already an initialization existing for the infoprovider.
What happens is, when you have an initialization existing and you run a full load (without specifying it as a full repair), then that reuest will not activate because of this initialization.
Hence you have to you to use a full 'repair' instead of full load so that it does not affect your existing init.
Full loads/full repairs/inits gets it data from set up tables (that you load according to your selection criteria)
Hope this helps.
Shoot any other questions you may have
1) - 2) You will get the 10,000 from initial setup. The setup tables are only filled when you execute the fill. The delta data has nothing to do with the setup tables.
3) You do not need to fill the setup tables to perform an Init. There is an option on the infopackage to perform the Init without data (this will only set up the delta queues). However, you can use the setup tables to perform a "Full load". Use the "Repair" option on the infopackage to extract data from the setup tables without disturbing the delta process.
Does this help?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello Geo,
Thanks for your answer...... That means each time If I run a Init load today (according to the above example), I will only get 10,000 and I will lose all the 30,000 records as the delta will start bringing new/changed records from today......
Hence each time I do a INIT I need to run the setup table so that I will have all records. Am I right ?
User | Count |
---|---|
93 | |
10 | |
10 | |
9 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.