Safety Intervals
Safety limits are used to subtract the time intervals of the
data pointers. Safety intervals subtracts the data from the limit values. If
current pointer is 1000 and there are 500 more records in delta and each time
there will be 500 delta records.
Lower Limit
If we add lower limit 100 , first run, we get data from 901
– 1500 instead of 1001-1500 first time, then again for the next time we get
1401 – 2000 instead of 15001 – 2000. There will be always repition of data.
Here 901 to 1000 and also 1401-1500 are extracted twice.
-
So whenever lower limit is set, we should always
load to DSO first and then to Infocube. Bcos of duplicates.
Upper Limit
If upper limit 100, we get records 1001 – 1400 instead of
1001-1500 and then, 14001 – 1900 instead of 1500 – 2000.
-
We do not get any duplicate records but we will
have lag in records.
Scenario
For say, a transaction take 2 minutes to get posted, and the
transaction was created at 1:00 PM and the delta runs at 1:01. Technically it
should pick the transaction values in delta. But as we said, it takes 2 minutes
to post the entry to the table, so it transaction value will only appear in the
table at 1:02 – hence not appearing in the delta which ran at 1:01. So if we
put upper limit at 2 minutes, at 1:01 it will fetch delta till 12:59 and in the
mean time 1:02 will be posted to the next delta.
Deltas methods are ‘Additive
Delta’ and ‘New Status for changed Records’ which is After Image..
