DSXchange

If Sainath has ten characters (guaranteed) from the input stream, then concatenation ("00000" : InLink.C10column) is the most efficient algorithm.

The ? you see here is not Char(63). Rather, it's the Unicode special representation of "unmappable character (from memory, somewhere near UniChar(0xFB)). Each worksheet is a separate table. You should, therefore, be able to process each with a separate job (clone), but with a different NLS character...

Since there is a virtual Data Set between your Transformer stage and your Sequential File stage (you can see this in the generated OSH), if there is a bug or a design problem it's within the Sequential File stage. Can you thoroughly check the metadata in the Sequential File stage and its input link,...

Null field value must be precisely the same length as non-null data if the file is fixed-width format. Null field length allows you to specify how long the representation of null is in a variable-width format column. I believe they are mutually incompatible. Click on Help with the Edit Column Metada...

Have you tried doing a lookup with NULL values in relevant foreign key columns to see whether that exists, and only then transmitting the row to update the table?

It is possible to prove that A RIGHT OUTER JOIN B ON condition is equivalent to B LEFT OUTER JOIN A ON condition for an equijoin, though I don't propose to provide that proof here. Stream input from table B and perform regular lookups against table A (or a hashed file populated from table A). Do not...

You can not safely delete any of these. If they are "taking up too much room" delete jobs that are not longer required. RT_STATUSnnn is small, and contains the status records for job number nnn. RT_CONFIGnnn is small, and contains the run-time configuration information for job number nnn. RT_BPnnn ...

Ken outlined the advantage of a hashed file. As to whether ODBC can avoid regenerating the result set, it really depends on the database server, and is not under control of DataStage. The technique used is "prepared SQL", where a query containing parameter markers is sent by DataStage to the databas...

Try choosing the input column name from the expression editor so that your input column name is fully qualified with the input link name.

An integer. A command line argument. Beyond that I have no idea. Examining the script may shed some light; asking your support provider may also prove fruitful, or may result in the response that you don't need to know. I very much doubt that it is the cause of "access denied" or "unable to fork", w...

Lookups in server jobs are achieved by painting one or more reference input links into a Transformer stage and supplying them from passive stage types that support "get by key" functionality (for example NOT Sequential File stage). You supply an expression in the Transformer stage that is used as th...

Not sure what you mean. The results from a lookup, no matter what stage type is on the other end of a reference input link, is a row containing either data or nulls. (From an ODBC stage or a UV stage it might be more than one such row.) Columns from these rows can be directed onto the output of the ...

The error message indicates that it is the T30FILE setting that needs to be fixed. You are not hitting the "sub-directories in a directory" limit - that would generate a rather different error message. The problem is not being caused by running both parallel and server jobs; it's just the total numb...

Open the Columns grid, scroll right and find where you can set the "column contains line terminators" rule. This should allow you to handle any intermediate line terminator characters in the data.

A routine that reads the mapping file then processes the data file seems to be the most viable solution. The routine can then create another file containing delimited data with column headings. You could then import the table definition from the new sequential file and design a job on that basis. To...