DSXchange

CWD is always the project. It's dangerous to not fully qualify all scripts and paths.

They are gone. Retreiving from a file system level backup is so perilous you're likely to corrupt your entire project. Job design information is stored as rows within an internal file, those rows have been deleted. Export your development projects daily with a date in the filename. You never know wh...

Is there any way we do this automatic. What I mean is, we will have both TEST and PROD information Parameter Values saved in the ETL and based on where we are running the appropriate (TEST or PROD) value are initialized? Any help appreciated. Intelligent job control, stored process metadata, and co...

Back to Premium Content... If you're doing something like a daily rebuild of a datamart, truncation or reorg methods are optimal. No matter what you're doing, you have to fit within rollback. Dave's original question is about removing 1/65th of the data weekly and adding 1/65th more. Could you justi...

I unmarked the Premium Content so that everyone can see the original post replies. The issue is what can be done within ETL that doesn't involve DBAs. Often we have to do things within ETL that are part of an automated load cycle. If you design a solution that periodically breaks because you've exce...

Because hashed files are not sequential files. Sequential files use exactly the number of characters in each row, plus either a LF or a CR/LF for end of row terminators. You can take the average number of characters per row (plus the end of row terminators) and multiply that by the expected number o...

Go look here: 'C:\Ascential\DataStage\Projects\Alaris_Prod/Hashed_Filename/DATA.30' Look at the size of the DATA.30 and OVER.30 files. If either one exceeds 2.2GB then your problem is 32BIT and you should create the hashed file as 64BIT. There is no datatyping, therefore, the metadata within DataSta...

If you mean column delimiter, then you could try using the pipe '|' character as the delimiter, and ignore the columns with just the decimals. Treat them as marker fields, just ignore them.

Sounds like you need to write a DS Function that will do your specific string parsing. Useful DS BASIC string functions are INDEX, FIELD, SUBSTR, LEFT, RIGHT, and COUNT.

You can look up invocation in the dictionary. In the DS application, it means an ad-hoc copy of a job created at runtime. You can read the manual for more details.

In all programming languages, a function derives a result, whereas a routine does "stuff". Functions always return a result, routines always return a status. Sometimes a routine's status is a result of a calculation. A situation where you would use a Function is when a value has a special calculatio...

The Director scheduler prompts for parameters and run limits. Just go to the other tab.

The pain, however, is to specify the param names on the canvas. Apples and toaster ovens. You had pre-existing jobs with literal values that needed substitution to parameters. I never have that situation because a parameterized architecture is a fundamental first step in the ETL design. If I'm goin...

There is a difference between blank and NULL. NULL is a physical character on the ASCII chart, blank is a zero length string. You can NEVER, in SQL or DS BASIC, compare ANYTHING to a NULL. You can't do math on NULL values either. That's why there's a specific function in DS BASIC and SQL to handle t...

If using Sequencers or custom job control, you set the limit then. If using Director, the default for the project shows up in the dialog box. It's your responsibility to set the the run limit on warnings in all cases.