DSXchange

Do a "search for all terms" on packed file.
This will retrieve many posts that will assist in answering your question.

First you will have to see whether the select or the insert is taking more time. Pull the records to a sequential file stage and make a note of the performance. Do it for just 100k records. Then load from seq. file stage to the target. Note the performance level again. This will give you an idea on ...

vsi wrote:a) Here I dont have or cant check for count(key)=1.

I understand that you cannot do a count on key as one doesnt exist now. But in order to to find out if duplicates exist or not, you need to do a count on the two columns.

Protected means read only. Cannot be edited.

Fixed width file and a delimitted file are two different formats. A fixed width cannot be delimitted and a delimitted cannot be fixed width. If you want the length of the columns from then source to be preserved and have a delimiter then do as Rohit suggested, but dont call it fixed width.

Are you sending in milliseconds. If you are then dont. SQLServer will stick .000 at the end.

Yes, but in an event of a rejection, warnings are generated. If you dont see any warnings then as Craig mentioned, maybe you are loading it somewhere else. Like say mabye in one environment and while hitting view data the default parameter might be pointing to someother environment. Possible ...?

And when you say duplicates not found, you mean duplicates on Col1 and Col2 for step 1
Col1 and Col3 for step 2
and Col2 and Col3 for step 3, Right?
Or is it completely identical rows as shown by your sample data?

Only your conditions have changed. Cant you make changes to the conditions in your other post and use that?
For others who are wondering what post i am talking about, its here
http://www.dsxchange.com/viewtopic.php?t=101864

Load the source in the same database as the lookup table and do a database level join.

Offcourse you can. Break the query into two. Query1: select B.Region,sum(B.Sal)As Sal from table B group by Region Since this query is at the Region level, it will retrieve lesser amount of rows as compared to the second query, load this into hashed file, keyed on Region. Query2: select A.Re...

Its pronounced "Dun Lairy" , Right ?

Here, read Kim's reply in this post
http://www.dsxchange.com/viewtopic.php?t=94025

In his second para, he has mentioned GenHtml will get you what you want.

You guyz are having too much fun. Way to much fun.
Enjoy

What is the format of date in your mainframe file. Can you give us a snapshot of sample data?