"Sreenath" <Sreenath@discussions.microsoft.com> wrote in message news:F184C98A-2AEC-4C73-9CB8-E0586D4D9ED1@microsoft.com...
> hi,
> I have flat file with atleast 500,000 rows with fixed fields. But, the rows
> were divided into 5 different types of rows. First 3 columns are same, 4th
> column tells us what type of row it is, in all rows and each row length is
> around 80. The remaining columns are different based on the row type.
>
> My question is, Is it better to dump everything into one temp table using
> bulk insert task and then write a stored procedure to separate the data into
> different tables based on the 4th column ?or Write 5 different datapump tasks
> and for each task based on the column type insert into a table.
> Looks like the first idea is good but I am worrying about the cursor memory
> as it has to hold 500,000 rows of data and each row contains 5 fields.
>
> Thanks in Advance...
> Sreenath

"Sreenath" <Sreenath@discussions.microsoft.com> wrote in message news:<F184C98A-2AEC-4C73-9CB8-E0586D4D9ED1@microsoft.com>...
> hi,
> I have flat file with atleast 500,000 rows with fixed fields. But, the rows
> were divided into 5 different types of rows. First 3 columns are same, 4th
> column tells us what type of row it is, in all rows and each row length is
> around 80. The remaining columns are different based on the row type.
>
> My question is, Is it better to dump everything into one temp table using
> bulk insert task and then write a stored procedure to separate the data into
> different tables based on the 4th column ?or Write 5 different datapump tasks
> and for each task based on the column type insert into a table.
> Looks like the first idea is good but I am worrying about the cursor memory
> as it has to hold 500,000 rows of data and each row contains 5 fields.
>
> Thanks in Advance...