DSXchange

Some useful functions for conversion (find them in on-line Help or in the BASIC manual) are: Convert() character-by-character conversion Ereplace() substring-by-substring conversion FieldStore() delimited substring replacement Is the "invalid property value" message being generated by DataStage or b...

Hi, I resolved this. I used ConvDateToYYYYMMDDTimeOracle(Oconv(Date(), "D-DMY[2,A3,2]")) But now, I have another issue. Column Flag has three values called YES,NO,NA I want to update all YES rows to NO I came to know to update a column and with where condition on the same column not possible. need t...

There is no sense of "appending" to a hashed file. The physical location of each record is determined by its key value. You can have as many inputs as you like into a hashed file (up to the design limit of DataStage, which I believe is 128 or 512 links per stage). However, note that the behaviour of...

1. Sure. You may need to watch out for duplicates. If two data streams write to the same hash key then one will overwrite the other. 2. If you make the lookup follow the write then it works fine. You may need to turn off write cache. DataStage does some buffering on hash files which can make results...

Hi There, I have a couple of questions. 1. Can I append to the same Hash file in the same job using two input files ? 2. In the same job can I insert and do a look up against the same Hash without creating an initialization job ? Can someone give me any ideas. Thanks. *******************************...

You can implement using user-defined SQL and the SYSDATE constant, or you can generate the system date in the desired format using the Date() function (or @DATE system variable, which is set at the beginning of, and remains constant for, the duration of the job). For example: Oconv(Date(), "D-DMY[2,...

Hi, I have two issues. I need to load sysdate in one oracle date column.How to implement in datastage job. I assume oracle date format is DD-MON-YY second one is in source file, the date format is dd-mon-yy, and how to load this data into oracle date column. WISH U HAPPY NEW YEAR Ramesh

Thanks for your comments, David and Alan, (I only get the daily digest of msgs, so I didnt see them immediately when first sent out). (David, No apology needed for misspelling my name; however, I have clarified something else in my signature line!) To your comments, thanks to you both for spending t...

David, Thanks for the clarification on the deliberate processing of the same natural keys. The requirement is that they need to appear in the correct chronological sequence in the input file. Ive tended to use the effective date as part of the key in order to keep things in time sequence. Alan > > F...

Alan, Tracy is deliberately trying to process the multiple input records with the same natural key to track changes and deliberately doing lookups to the same table she is updating. I have used this design myself many times before. The only catch is that you have to make sure that the updates are co...

Tracey, I understand what you are getting at. I use this approach with Oracle in quite a few jobs but committing is an area where Oracle and SQL Server differ a bit. Also, the ORAOCI stage handles commits a little differently to the ODBC stage. You mention trying the "auto-commit" setting. I think t...

Tracy, The problem stems from having multiple records in your input file with the same natural key, and that youre reading from and writing to the same table. Do these records with the same natural key appear as a group or are they scattered throughout the input file? You may want to enforce uniquen...

Hi, I have a collection of several jobs, all similarly designed, that feed flat file comma delimited input into the dimension tables of a data warehouse (in SQL Server 2000). The basic job flow is: Sequential Input stage >> Check Foreign Keys transformer (do hashed file and ODBC lookups to validate ...

>From what I recall, you cant do a wholesale conversion from EBCDIC to ASCII if there are packed fields in the record (using something like dd) as this porks up the packed fields. You either need to use a utility that allows the mapping of various portions of the record, some of which get converted ...

Yes, its somewhat indirect; the job control code in the sequencer itself is locked against edit. However, you can "Select All" by dragging, copy the generated code, and paste it into a new server job. Then you can edit it. Apologies for not making this clear earlier. I have just confirmed (in DS5.1 ...