DSXchange

Even i built a test job yesterday with DSJobStartTimestamp, but i never hit view data. I went at the unix level and did a cat. It wasnt reading it due to the space between the Date and Time. I just ran the job again and hit view data, and yes you are right, it is able to view data. But i would still...

Ouch, totally bypassed that. Need the column name huh?

Well then, before sending it to the target for insert, you need to flag one of them as inactive and filter it out.
Like in one transformer flag it and in the second transformer constraint it.

Are you trying to open the file via seq. file stage?
If yes then it wont open it because you have a macro in the name. YOu need to go at the unix level and do a cat of the file.
And also, just for your own satisfaction, at the unix level type
touch 22:22 and see that it does get created.

Even if its not there, something like that can be built. Like maybe have a lookup set with the character value and its weight. Get that value and then sort it based on the weight.

O well.
You need to go bang the administrators head now
Na, i am just kidding, be polite.

Well, i dont think a function exists. In server there is Compare(), but that wont even work over here because the number of columns is more than 2. As Kumar said, you can do this via if-then-else statements (lots of them for 10 columns). I would go for the case statement in sql. I know its labouriou...

Could you kindly share the solution with us, so that future posters, encountering a similar problem, will know what to do.

It may be because of the timestamp format eg "14:44:36" as unix has problem creating/reading file having ":" in the name Says who. The problem is not with the charcter ":" but instead a space. The macro "DSJobStartTimestamp" creates a date part and a time part with a space between the date and time...

Yea, just pass that sql command from your database stage. Even that is called "Implementing in DataStage".

Only if you have another column that specifies the sort order, or the character weight.

Basically in other words, you need to update your target table with only the first five records from the input. or there is more to it, because that sounds a bit to easy. Take you input (file, table), constraint it only to the first five records and update your table. I know there is more to it than...

Actually Kris, i just looked at my code again. If you want it at a per table basis then my code will work. But if you want the max date from three columns at the row level then you need a case statement. Use this select (case when date1 >= date2 and date1 >= date3 then date1 when date2 >= date3 ...

What is the size of your source file?
Why exactly are you using CFF stage?
Did you try accessing the file with seq. file stage? What happens when you do, dont worry about junk data, right now confirm if you are able to view it.
Also, make sure you are sourcing the correct file.

Kris, do it in the sql itself