Server Jobs - record processing

Manikandan · Post by **Manikandan** » Thu Sep 15, 2005 7:25 am

I have 10 records with 2 columns each in my input sequential file like,

A1 B1
A2 B2
A3 B3
A4 B1
A5 B5

I want to generate Surrogate Key for each of the column (should result in 09 surrogate key) using hash file lookup. According to my understanding when we are using a dynamic hash file records get cached for every record in the memory, however it is not happening so.... in the same input file when I have duplicate col2 for some records, the lookup is failing and I could record the same with a flag.

Wish to know, even in Server Jobs records are not processed row by row but in bunch.

Thanks,
Manikandan

ray.wurlod · Post by **ray.wurlod** » Thu Sep 15, 2005 4:14 pm

"Wish to know, even in Server Jobs records are not processed row by row but in bunch."

If row buffering is disabled, row by row. If row buffering is enabled and you have active-to-active stage links, then rows are buffered ("bunches"?).

Manikandan · Post by **Manikandan** » Fri Sep 16, 2005 12:54 am

ray.wurlod wrote:"Wish to know, even in Server Jobs records are not processed row by row but in bunch."

If row buffering is disabled, row by row. If row buffering is enabled and you have active-to-active stage links, then rows are buffered ("bunches"?).

Thanks Ray, but if row buffering is disabled again we may get caught with performance issue. For now I have handled the case by putting in an intermediate Hash file and avoiding duplicate in target.