DSXchange

I had to go back and check because I was certain that the DB2 stage had sparse and normal lookups. It does, but only the DB2 Enterprise stage. Do you have that stage available?

Oh, one other thing. If one of your keys is nullable and contains a null value, the lookup will fail.

Strange. Did you restart DataStage as root or as your dsadm user? Also, check your $DSHOME/dsenv file to see if it contains a ulimit command.

I am confused from your post if you are looking to truncate to 2 decimal places while converting from float to integer or converting the string. The first could be done by math, i.e. INT(In.Col*100)/100; the latter by In.Col[1,INDEX(In.Col,'.',1)+2]

arsh14, welcome to DSXchange. I am not familiar with the rolllike function and can guess at what you wish to do but cannot be certain. Perhaps you could explain the logic in text form and you will probably get more accurate and quicker responses to your question.

Hello and welcome to DSXchange!

The part of your error message that reads

...Error when checking composite operator: cxx -O -...

should point you in the correct direction - your c++ compiler is not installed correctly.

Would you be looking for the default value in the job definition or the historical values actually used at runtime?

I am working with columns containing similar values. In my case, I added a filter of "tr '/000' '?'" to replace the 0x000 nulls with question marks in the string.

n.b. 0x000 is NUL, not an SQL NULL!

Can you change the lookup to a sparse lookup, that way you will get SQL generated by DataStage. If that fails, then copy the SQL into your DB2 query tool and plug in some actual data values to see if the lookup fails there as well.

Do . Is that clear enough?

In parallel jobs you can use two basic methods: a) cause failed lookups to go to a reject link, then process those rows accordingly b) all columns returned on failed lookupwill be null if "continue" is set. If the column is not nullable the default value is assigned. In a subsequent transform stage ...

If you want to get the duplicates, I think you will need to use a UV stage (or ODBC) to return all those duplicates instead of using a hashed file type stage.

The first thing you need to do is figure out what you want to do. Which of the 10 returned records do you wish to use? Will any of the rows suffice? Do you wish to work on all 10 records?

please re-read my post. I think it is the "root" user causing the issue. If a process has ulimits and spawns another, then the new one cannot have a higher ulimit (even if the settings are unlimited) than the parent.

set up DB2 on the server which is running DataStage so that you can issue "DB2 connect ..." to the DB2 instance. Once you can do that, you will be able to do the same thing from DataStage.

Yet your error message shows that you are converting this string to an integer in your sequential file stage (which one - you have 3 of them in your flows). On the 1st sequential write stage did you change the default file settings, particularly the quote characters and attributes?