Hi all,
Can you please comment on the implementation/solutions of the below requirements-->
this is our current process-->
1)we get some data files every day from source which we use to load through tpump utility in to a work table .
2) If some records gets inserted into error table of tpump then teradata automatically aborts that tpump operation.
3) if there is no record in error table and tpump is successful then we insert -select all the records from work table into the main table.
New requirement -->
Even If we some errorneous records are present in the data file of a particular like date- time, not null issues then
a) we dont want tpump operation to be aborted based on that but we would like to capture those records into the separate error table having the complete
copy of the errorneous record( unlike E1,E2 tables of tpump where it just captures the sequence no. of the erroneous record)
so that these are available for the investigation in future.
b)And all the good records to be processed into the work table with out any abort in tpump.
Solution-->
1)We are thinking abt increasing the ERRLIMIT feature in tpump but still we would need to capture the complete copy of the errorneous records
which seems not to be possible here.
2) the other solution may be if we define all the columns in the work table of tpump as varchar and nullable so that it doesnt leads to rejection
in tpump process.. and we get all reords in work table.
from the work table we can filter the good records and errorneous records by checking the not nullability and date -time validations
during update/Insert operations in a bteq script
Please let me know ur suggestions on the above..
thanks in advance!
cheers!
Nishant