Search found 53125 matches

by ray.wurlod
Mon Jul 24, 2006 9:54 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: TimeStamp problem with DB2UDB Stage
Replies: 6
Views: 1459

Even ICONV(L1.END_DT,"D-YMD[4,2,2]"):" ":ICONV(L1.END_DT,"MTHS") will not give you what you want. What is the data type of END_DT ? But there is no internal format of a timestamp; so that ICONV() questions are irrelevant. You must construct a timestamp as a string of 19 or more characters, as expect...
by ray.wurlod
Mon Jul 24, 2006 9:51 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: DSR.ADMIN: Failed to write to project file.
Replies: 10
Views: 7209

What other user? Who issued the chown command? Is there no management control at your site?!!
by ray.wurlod
Mon Jul 24, 2006 9:50 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: rebuild Index
Replies: 3
Views: 1174

... and only if you have exclusive access!

If you don't you can leave the repository database in a worse state than when you issued the command. Luckily this can be corrected by another reindexing command, this time with exclusive access to the project.
by ray.wurlod
Mon Jul 24, 2006 9:47 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Append in Oracle Bulk loading
Replies: 13
Views: 2974

I've always found it better to use a Sequential File stage, and a custom-crafted CTL file. The CTL file might have been prototyped using the stage, but I just can't see the value in rewriting the same CTL file every time. There are two places throughput will improve; you can tune the CTL file (for e...
by ray.wurlod
Mon Jul 24, 2006 9:44 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Issue with UtilitySeqEmailAndAbort
Replies: 4
Views: 2396

Is it reproducible, or did it only happen once? If the latter, the mail client or server might have been off having a smoke or something at the time. If it's reproducible, try simplifying what you are doing as much as possible, then add complexitiy a piece at a time till it breaks, so that you can i...
by ray.wurlod
Mon Jul 24, 2006 9:42 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Operator terminated abnormally: received signal SIGSEGV
Replies: 7
Views: 27154

SIGSEGV is the UNIX signal meaning "segmentation violation", an attempt to access a memory address not owned by the process. Somewhere, something has run out of memory.

Curiously you are not the first to encounter this. Did you search the forum for a match on your error text?
by ray.wurlod
Mon Jul 24, 2006 9:39 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Trim in FILTER stage
Replies: 7
Views: 6618

One stage, one task. Fundamental tenet of parallel job design. The Filter stage does not trim. You can trim using either a Modify stage or a Transformer stage. No other stage has a Trim() capability, unless you write your own custom stage. My advice is to use a Modify stage upstream of the Filter st...
by ray.wurlod
Mon Jul 24, 2006 4:27 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Variable-length record
Replies: 15
Views: 11546

Or just use FILLER to make sure every record has the same length.
by ray.wurlod
Mon Jul 24, 2006 4:25 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Using DataStage reported "Device" and "inode&
Replies: 3
Views: 1080

There's an executable $DSHOME/bin/UVfile that can help to find these "orphaned" hashed files. Use find to capture names, exec UVfile , then grep '(' to select only those with UniVerse file information in their reports. For example: find . -print -exec $DSHOME/bin/UVfile {} \; | grep '('
by ray.wurlod
Mon Jul 24, 2006 4:15 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Append in Oracle Bulk loading
Replies: 13
Views: 2974

You could use the stream editor sed for this task.
by ray.wurlod
Mon Jul 24, 2006 4:13 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Timestamp datatype
Replies: 7
Views: 1809

Easy. You supply a timestamp with microseconds.
There are no inherent data types in server jobs, so you can construct it as a string, and use an appropriate date picture for the TO_DATE function in the INSERT statement.
It's up to you where you get the microseconds from.
by ray.wurlod
Mon Jul 24, 2006 4:11 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Help req for setting proper ENV variable
Replies: 9
Views: 2472

ENV is a reserved word - it's the DataStage command for getting and setting environment variables. Use a different name.
by ray.wurlod
Mon Jul 24, 2006 4:09 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Issue with UtilitySeqEmailAndAbort
Replies: 4
Views: 2396

You can get this message on any failure to send, for example if the mail client is not running. That is, the problem is not necessarily in DataStage; "they" are trying to be helpful with diagnostic suggestions, too helpful perhaps, since it can be misleading. Of course, it could be that the job para...
by ray.wurlod
Mon Jul 24, 2006 4:06 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Datastage releases
Replies: 1
Views: 844

Version 1.0 came out in November 1997. Therefore no-one has ten years' experience with the product unless they were working for the vendor at the time. This does not stop recruiters requiring ten years' experience. As to the rest, that would require some research, particularly as to dates. You will ...
by ray.wurlod
Mon Jul 24, 2006 4:00 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Using DataStage reported "Device" and "inode&
Replies: 3
Views: 1080

Run ACCOUNT.FILE.STATS ALL LOCAL in the project on a regular basis. This updates a hashed file called STAT.FILE in the project. Apart from having gathered file tuning statistics, STAT.FILE contains the device, inode and file name for everything accessible from that account's VOC file with type "F" o...