November 8, 2011 at 2:54 am
I agree with Koen, a well designed ETL Packages wether it be in SSIS or any of the other tools should have logic in there that makes the inserts Atomic (all or nothing)
This can be done via transactions, or if they are not supported then there should be a 'clean-up' task that removes all the rows that have been inserted that are related to the failed run. This should happen as part of the ETL package and should not require re-running anything.
November 8, 2011 at 7:14 am
SQL_Athmi (6/6/2011)
Difference Between the MERGE and MERGE JOINMERGE - similar to JOIN operation. Only joins only two inputs. If there are more than two inputs UNIONALL is used
MERGE JOIN - select the type of JOIN (LEFT, INNER, FULL) you wish to make.
Similarities between these two:
1) Both needs sorted input.
2) Same i/p type (Int, char)
Kinda an old answer, but still, I feel this needs correcting.
MERGE is not similar to JOIN.
MERGE is essentially the same thing as UNION ALL, except that it has a couple of extra requirements. MERGE can only be used on sorted datasets, and can only be used with two datasets.
I've searched around and I can't really find any justification for actually using MERGE instead of UNION ALL, so for all intents and purposes, unless you know your data sets are already sorted, you may as well just use UNION ALL 😛
MERGE JOIN on the other hand, is a JOIN operation, and can be either an INNER or an OUTER join.
November 9, 2011 at 2:45 am
I have asked peoples who are interested in reading FQA's the most frequent asnwer i heared is to get the good job by memorizing the Q & A , but few of them told to improve there presentation skills & to understand the way of giving the answer to question they will go for reading FQA's and good articles.
Viewing 3 posts - 46 through 47 (of 47 total)
You must be logged in to reply to this topic. Login to reply