DSXchange

Rajesh, take a look at the documentation on the "merge" stage in DataStage Server edition, you will find that this addresses your question, at least partially. You might also think about loading the table to a hashed file and doing a lookup on that (you can load it to memory if it fits to further in...

DataStage BASIC does not allocate contiguous space for strings, it uses a very nifty internal mechanism to store a string as a linked list; and each element in this construct has a different length (I think it's based on a Fibonacci series). This means that you won't need to malloc the whole chunk o...

tail does it by positioning to the end and reading chunk of {buffer size} from there. You could do it in BASIC as well, but I haven't measured how good the performance is with SEEK AND BLKREADs

Since a sequential file on UNIX and Windows is never indexed; you will need to make at least one full pass through the file to get this information, even if your file is fixed length. To limit it to one full pass, read each line and keep count, storing the {n} last lines in a list. If your number {n...

Yes, the value is correct; the comment in the documentation is meant to avoid programmers checking the validity of the variable using integer conditions (i.e. FileHandle < 1). You should use the documented methods in the routine description to detect if the call was successful.

You can get a pretty good list by entering the command "HELP"; but of the hundreds of commands you will get only a few will be of use or helpful for you. The main concept behind DataStage is that you can use the graphical client software instead of the text commands. You cannot "enter DataStage" fro...

This routine is not meant to be called for each row in a transform stage; it will show in red in a Transform because it is not defined in the list of valid functions. DSAttachJob and the other DS... functions documented in the BASIC manual are meant to be used from within your own routines.

I can't tell you why this routine isn't working the way you want. If it is part of a derivation or a stage variable or a constraint in a transform stage then it will get executed for each row of data, no exceptions. You will need to use the debugger and/or some logging statements in your routine to ...

The DataStage engine is still quite close to what IBM now own and sell as UniVerse. You can download the manuals from IBM's web pages and one of the utilities documented is the UCI (UniVerse Call Interface) and Unibjects which describes the API you can use to call UniVerse from a language such as C+...

Force a length by using a derivation like In.ColName[1,3] to limit it to 3 chars regardless of source length. Does the error still occur? And I am quite confused by you saying that in some cases you found an extra character but the file was perfect. It is never wise to mix CHAR and VARCHAR types and...

From you description the 1st part of your question isn't answerable; if you call your routine 3 times it should get executed 3 times. You need to revisit your transformer code to see what is happening. You can position to the end of a sequential file by issuing a SEEK F2, 0, 2 ELSE CALL DSLogFatal&#...

Urs,

could you answer Craig's question? Also, when the job runs successfully the other times does it process more than 2000 rows and if you look at the log entry does it read limits of 0/50 or 2000/50?

You either need to compile or reset your Job C

Could you add a dummy column to your output with the derivation of LEN(PXI_OUT_1_Lnk.MED_NON_COVD_REAS_CD ) to make 100% certain that this string is of length 3 and doesn't contain padded spaces and/or non-displayable characters? The length limitation given in DataStage server jobs will not truncate...

This means that you cannot run DSJobReport at Version 5.2; it contains functionality not supported at this version.