thanks for the reply rudy.
now if i got a million rows in a table and need to select the data on some criteria, return these records to the calling application
and after this selection need to update the records which were selected above.
I do it in transaction like:
begin transaction set1
select records from Table_name
where some criteria
update table_name set records where some criteria
now this is a multi row operation (if im not wrong)
how can this be converted to a single row operation ...
someone suggested that single row operation will me more scallable
scalablein terms like if some operation is to be done on some chunk of data (say about 5 million rows) then for a multi row operation each time the whole table is scanned even those records which are already been operated on, now what I am looking for is a stop start type solution like I have 10 rows , I execute a sp which performs some operation on these 10 rows, so the next time when the same sp is run, it wont read these records staright away start off with the 11th record. some thing like a pointer ...
I m sorry if im making it confusing to understand...