DSXchange

Its how the data is read from source. If its extracted as character, it has to be converted into Decimal and as the process the comma need to be excluded.
Defaults tab in Job Property will also help in specifying a different Decimal separator other than .(period).

There are lot of ways to get it done. If you make a search you will find a lot of idea.
One simple option is to convert the dates into Julian dates and subtract. The idea is to convert the date into internal or integer format.
JulianDayFromDay() can be used for that.

You mean 'Modify' Stage? Its possible.
As noted Copy stage is more lighter that Modify stage.

Btw, space is not NULL. Where do you find the space character is getting changed as Null? Is the field a Varchar datatype?

I should better learn to phrase sentences clearly.

You are appending a flag to a file. Like
Cat 'Flag' >> File.
You can use tail command to read the last line in CommandActivity and use the CommandOutput to pass the value as parameter to the job.

If you think its resolves your issue, you can mark the topic as resolved.

I don't get you full business logic on why is it all been done.
Just be aware that, if your input have more that 1 row with the same key value say 1001, you will end up in inserting only the latest one. Where I guess you need to insert both of them.

Have you tried the option "Replace Existing Row completely" with single filed marked as Key in Datastage.
Observer the SQL generated, a DELETE statement will be issued first with the condition, Where Col1 = :1 (Single field).
Later Insert statement.

Even Server jobs has CFF stage as plugin.

As recommended in document, go for sparse in 1:100 ratio. That is if you have input records 100 times less that the lookup data. In other way, very large data to be loaded into physical memory where as fewer times of hits are required to the lookup table.

What you expect the output should be for this case? If there is any record with 1001, it should be replaced with these two records? And by the way, the database keys has not been changed, only are just marking the key in datastage as one key. But to be more comfortable, you can do a lookup to the ta...

Hi and welcome Aboard :D !!!!

Have you searched for "Code=99". There a quite a few hist for this.
Like this

You need to replace based on single column, am I right?
Mark that single column alone as key in Datastage.

Probably the other way, Varchar to Char.