DSXchange

I just realized that it is using "/TMP", this would be the value for the TEMP directory as specified in the uvconfig file. Try creating "C:\TMP" (use the correct case, just in case) and see if it works then. Either that, or change your uvconfig to point to a correct location.

DataStage has no knowledge of "HIGH VALUE" - it is a COBOL concept. In most cobol implementations a PIC(X) high value is 0xFF but I seem to recall that there are differences depending on the COBOL implementation. ANSII Standard would be 0xFF. I can't recall what is used for COMP-n data types, I don'...

You will have to write a loop in this case since you have 2 values, the start and end. This would best be done with a wrapper stage.

Ok, does ""E:/Ascential/DataStage/Scratch"" exist and does your userid have RWD access to it?

Using " " (22 spaces) is the correct way to do it. What is the exact error you get when you use this value?

"high" and "low" values are a COBOL concept - is this data going back to COBOL? You will need to check what the EBCDIC "HIGH VALUE" is and either use that value in a binary field, or use the appropriate ASCII mapping for a non-binary one so that when the file is converted back to EBCDIC those values...

What have you tried so far and what errors did you get?

...that will convert the NULL to Blanks and trying to work it out on the problem... You cannot convert NULLs to Blanks in an integer field (at least not in a transform stage)! Let us keep the test program simple, just a source Oracle stage to read 1 column that is a width 1 integer, then write to a...

You have now re-written your file without delimiters and quotes, so it will look like your original file except that all occurrences of '\"' are now '""'. Now read it with your original 6 column definition.

The complexity of a join doesn't change depending upon data source type - it is a matter of your join logic.

No, I have not seen this error before. And we are trying to help you get around this problem - if there were a known solution someone would have posted it already. Can you write a simple test job using just one of these integer columns and no others. Have you checked the DDL for *both* your source a...

How did the pipe character get into the output file path name? A "named pipe" is something very different. Try it with a normal sequential file first.

I hope that your F:\ drive is local to the server. What does your APT_CONFIG_FILE read for the scratch path?

Read the file as just one big column with no quotes. Use "EREPLACE(In.BigColumn,'\"','""') to convert all \" to "" (the correct representation) and write that back to a file with no quotes. You can then use this new file as your input and DataStage will correctly convert the "" to single " in the st...

You APT_CONFIG_FILE specifies a UNIX pathname for the resource scratchdisk which doesn't exist on your windows machine.