DSXchange

It would be stored in the RT_STATUSnnn table for each individual job where 'nnn' is the internal job number. From what I recall. You may be able to do an exact search here and turn up more details, I don't have any with me at the moment.

Just a quick question right now - what exact version of DataStage are you running?

Good to know. I've just always preferred being... explicit.

You don't need two stage variables, just the one will do if you add the 'else' that is missing from the initial suggestion: svABC: If XYZ = 'S' Then NextSurrogateKey() else svABC This will preserve the value for all non-'S' rows. Of course, you will also have to worry about your partitioning adverse...

So DB2 fails because it is comparing "XX " to "XX" is what you are saying? If that's the case, then yes you do need to trim that source field but you need to ensure your target for the trim() function is a varchar. Trimming a char to a char is a waste of time as the datatype just...

Well... you are transforming a date to a timestamp so I really don't see a way to do that without adding something for the time, specifically the zeroes you need. You can try using a TO_DATE() in your source SQL directly into a Timestamp field when selecting that field: TO_DATE(YOUR_FIELD, 'YYYYMMDD...

Out of curiousity, what do you consider to be a 'direct way'? To me, that is one. Do you perhaps mean an implicit conversion? If so, what is your source for this - a flat file, another database?

devsonali wrote:I can append a time and then use this function but i don t want to do it .

Why not? Append a zero time as part of the function call.

They've specified 8.1 in the Additional Info field. A recent discussion of the function and that format is here.

What 'this' is being argued as a reason to not use datasets?

First of all, it must be a supported compiler for your platform / version. And yes, you need to set environment variables / paths / etc so DataStage knows where it is and what options to use. And they are all documented. I randomly found that url for 8.1 with a search, make sure you look at the one ...

Or that.

Ah, true... so: 0 12 * * 1-5 /x/y/Scripts/Script.sh 2>&1 Would be more better. However, I'm curious about the '2-31' you have for day of month, did you specifically want to exclude the first day of the month if the 1st is a weekday? If so you may have to check the day inside the script and skip ...

Really not sure why that value would be an issue... but then I have no direct SQL Server experience. Is that your only issue, meaning other values load correctly and just something that starts with a hash ends up null?