DSXchange

The behaviour with nullable and not nullable varchar fields read from sequential files is both consistent and logical (as well as being documented). The different behaviour is caused by sequential files not containing NULLs. If you edit the attributes of your varchar() column in the sequential file ...

I know that you can do this at Version 8, but cannot recall if it is possible in versions before that.

We have encountered issues with Parameter Sets at V8 to which we received a client-side patch; but we never saw the "uncompiled" state problem.

You cannot modify input arguments. At the beginning of your job, add

TempFROM_MAILID = FROM_MAILID
TempTO_MAILID = TO_MAILID
TempSUBJECT = SUBJECT

and use the Temp... variables in your job.

What does the log file "/scratch/ora.930028.244413.0.log" contain?

What happens if you try

NAME2:String[100] = substring[1,100] (NAME1)

That is strange behaviour; you mean that you have "reject" enabled in your Oracle Enterprise stage and you have a reject link going out of the stage and you still have records that disappear on write - ending up neither in the table nor in the reject link?

If you search the net using Google you will find MS Outlook threads dealing with that issue and somewhere in there will be a description of the settings you need to change in Outlook in order to get rid of the empty attachment.

If you do a "diff" on the 2 files are they identical?

roopanwita - put in PRINT statements to see if you get valid values for your calls, i.e. print out FROM_MAILID, TO_MAILID, and the evaluation of the parameter to DSSendMail.

Since you cannot insert an SQL null into the columns, your second job needs to insert or update some value. Perhaps an empty string, or a string of "??????"s or something else - anything but the null.

You can do this in the job, but you need to explain While updating it is looking for values in all the 5 columns of a particular row and if it doesn't find a value in a sigle column it is rejecting the entire row without updating. How are you looking for values? In a transform stage? Where is the re...

BLOBs and CLOBs aren't supported. You can actually write to a LOB, but cannot read from it. If you know your maximum length you can write a workaround SQL query that uses CAST to move 4000 character substrings from the LOB into VarChar2(4000) columns. We just did this for LOBs that we know won't exc...

DataStage and DataBase metadata should always be identical. When it comes to nullability settings, the two need to be identical.

Then you need to either modify the definitions on your table or add empty values for the sparsely populated columns. The easier method is to make these columns nullable, that way if you don't specify them on an insert they won't generate an error, just a null value.