DSXchange

Check your metadata thoroughly. Get the player processes to dump the actual schemas on which they operate. Look for duplication of fields called A_I and B_I (which are adjacent in the culprit record schema).

Don't forget that the subsequent jobs will need to use a configuration file that is the same as, or compatible with, the configuration file used to write the Data Set. In particular all the nodes in the writer must be available in the reader.

Use None as the quote character in your target Sequential File stage. This is how you tell DataStage to use no quote characters surrounding strings.

There's no WriteSeq statement that would write anything to the flat file. If opened with OpenSeq, the statement to write a line is WriteSeq, not Write. If you want to use Write, open the parent directory with OpenPath, rather than opening the file itself with OpenSeq. Also, if opened with OpenSeq, y...

At some point you are going to need to SORT the data (at least for the second pass) to get a rank.
Then you need to implement your business rule for tied ranks. For example 1,2,2,4,...

What Arnd called "percentage rank" is simply "percentage".

Does it happen every time? Usually this error occurs because of a "blip" in the network, and is only seen rarely. If you can generate a reproducible case, involve your support provider.

If there should be exact copies and anything is "disjoint" then QualityStage is probably not the right tool, DataStage is. If, on the other hand, you need to allow slight variation (for example dates being different by one or two days, letter transposition in words, and so on) to be considered a mat...

Changing the specification risks invalidating the answer. There is no First or Last set function in the parallel Aggregator stage. Is that clear now? Your solution will, therefore, require both stage types. Use a "fork join" design; split the stream into two, run one stream through an Aggregator sta...

Why do you want U to explain? U does not log in all that often.

Welcome aboard.

The best solution to this is a lookup to a "time dimension" table or Data Set. Two columns in such a table are typically MonthFirst and MonthLast dates.

Keep going - you need more than the sum to get rank.

INSTEAD of Aggregator stage.

With NLS enabled, source characters that correspond to the dynamic array delimiter characters (Char(248) through Char(255)) should be re-mapped to the Unicode private use area. This is described in the DataStage NLS Guide. Therefore I don't understand how 0xFF is getting through the map on the bound...

Welcome aboard. There will be an appropriate NLS map for Chinese (you neglected to mention whether it's traditional or simplified characters), which is probably the one you ought to be using. On a Windows server try MS950 (traditional) or MS936 (simplified) for starters.

Again, this falls under the heading of testing - you need to set up conditions (such as a faulty input argument) and verify what your DataStage does. You have already observed that it had been configured differently from Craig's - who knows what other differences may obtain?