DSXchange

Another Transformer solution - maintain two stage variables to determine whether you've seen "<![CDATA[" (if not you're still in the header and you don't pass this row) and another to detect when you see "</xtd:Data>" (which means you're now in the trailer and you don't pass this or any more rows). ...

OK, not many. The easiest solution (in my opinion) for an arbitrary number of columns is in a server job, using the dynamic normalization capability of the Universe stage. Are you able to use server jobs at your site? If not then use a Sort stage to sort the data by key and to generate a key change ...

You have offered no indication as to what the actual "problem" is. What are you trying to achieve, what are the expected results and what are the actual results? Perhaps provide a row or three of sample data from the inputs and the output if you're convinced that the "problem" (whatever it is) is in...

Yes but the Manager client doesn't.

Therefore the answer is "no, not really".

You can create a script to use the dscc command, but it's a long and tedious process.

And why do you think posting the same question is likely to yield a different answer than a request for clarification on the original thread?

What data volume are we talking about here?

You can use attachments at IBM DeveloperWorks though that site is not as busy as this one - on the other hand more IBM people browse there (they are encouraged to do so).

Are your data partitioned on the primary key? If not, one partition may still be holding a lock while another partition is trying to acquire it, and vice versa (classic deadlock situation), particularly if your rows per transaction is greater than one.

That might be the case if you know what's in the repository. From memory a clean install on Windows gives an XMETA repository containing 847 tables. No-one, as far as I am aware, has taken the time to document their contents and inter-relationships and how these pertain to particular products, at le...

None, because there is no such thing. However, the ability to mark a column, or columns, as "key" can be useful in designating the possible future use of that column, or to indicate that that column has a unique property. Marking a field as key can also to indicate to the ODBC driver for text files ...

If you are on version 8, FORGET about using the local Repository tables alone.

Somewhere in the Information Server common repository there are additional references to these objects.

No-one has yet provided any ideas, scripts, tools, etc., for scanning that.

Depends. What is the data type of the output column? As I stated, one needs to know both to determine whether an implicit conversion can be done or whether to override the default rounding strategy or whether an explicit function is required.

Welcome aboard. What is the record delimiter character in the stage with which you are reading this file? If you don't want that to be the delimiter, change it to some character that does not appear in the data. Click on the Help key (or F1 key) for more information about options for entering this c...

What is the current data type that you want to round, and what is the data type of the field where you want to store the result. Knowledge of both of these things is essential before one can choose the correct function.

Funnily enough you need the Index() function. Index(InLink.TheString, "Index", 1) = 0 can be your constraint expression. You *could* use Matches but need to specify the wildcard characters correctly. Not(InLink.TheString Matches "...'Index'...")