DSXchange

Also, along with the null check, check for the length of the input argument. If its equal to zero, assume it to be null. Because if your input is a flat file, then ISNULL will always return a 0, but the Len() function will capture the length as 0.

So shall it be written
So shall it be done.

Actually thats what i meant to say, a Transform function. I dont know why i categorize everything as a 'function' Bound by habit i guess.

"For every action there is always and opposite and equal reaction.",
Stop acting like crap to it and it will stop behaving likewise.
Hey, i didnt say that, Newton did.
Jusk playing with you bro. I cant tell you much because it never happened with me.

Either the source or the hashed file has spaces in the key. Make sure you trim the keys while building the hashed file and also trim the keys during the lookup coming from the source.

MCN with OCONV extracts only numbers. You can also look at the function DIGITS(). It does the exact same thing.

Does the log show that the data got loaded or are you guessing that by looking at the link count?
Well, i guess all of your records are getting rejected due to one reason or another. Look into your log file.
OR
you are loading into one table and checking some other table.

Well, a few suggestions in writing the routine. Read in the entire description and take four positions at a time and check for "P/N ". If it doesnt match, store it in one variable which will be called your 'Restofdata' and if it matches, store it another variable untill you hit , Something like Desc...

Because your script is expecting two arguments and you have provided none in your command

If you are going to do that then make sure you have an annotation that specifies what the delimiter is, as it can easily be mistaken for a space.

Look into the CDC stage.

Ok, this is a little bit tricky. Before building the hashed file, add another column which is a surrogate key, a running number, i.e, 1,2,3..... and so on. Specify that as key and in the derivation of the lookup specify @INROWNUM so that each and every row has a hit. Define a stage variable to check...

You can also stick an IPC stage in between the transformer and the link collector stage to get rid of that error. This error occurs because both transformer and link collector are active stages and DataStage will not allow a direct link between them unless you have an IPC stage in between or have In...

use && in place of ; in Craig's code. This way the second command will only be executed if the first one is successful

rule: Only load the records where main_id = lookup_id and dt_record between dt_start and dt_end. But i thought it was just a single key as per your rule above. Well it doesnt matter how many keys you have. Just link the keys properly to the hashedfile in the transformer. Not(RefLink.NOTFOUND) will ...