DSXchange

the core dump is just an effect, when a program gets it's internals so messed up that it can't continue and aborts it will generate a core. I can say with confidence that if you look into almost any one of your project directories you will see a core file there. The error in this case is most likely...

"Workaround" implies a bug - in this case it's just a plain restriction and no, there is no workaround for that.

Could you make your VarChar() columns larger to accomodate the actual data length with a potential decimal place - and see if it works and what the data values are?

Can you show us rows 5,6, and 7?

Narayana,

the text you quoted is not HTML and not XML; it is unstructured text. You can use DataStage to read this and could parse it into variable length pairs of data before a semicolon (name) and after (value). This would work if the text ":" doesn't occur in the text.

If you have checked your master job properties and no longer have an after job routine call, please check your transform T1 and ensure that it doesn't have an after-stage routine call.

Bala,

you might not need to write a routine in order to do this. What is your 'rule' for determining whether a comma is a part of a string or a delimiter?

Me neither, but the docs state that it should work; I've tried a number of UV_SCHEMA definitions and decided to wait for Ray to post a response. I have 7.5.x as well.

Write a Transform with a derivation "EReplace(In.Column,'\\\',CHAR(010))" to replace all occurrences of '\\\' with a single LF. Write this to a sequential file or pipe with no formatting - no quotes or separators and use the UNIX style line termination. Then read the file and it will be in the forma...

You still have an after-job enabled :

Attempting to Cleanup after ABORT raised in stage CMACS3010..AfterJob

If you are reading from and writing to the same hashed file in the job and you need the "written" record to be "read" subsequently in the same stream then you cannot pre-load the file to memory and you need to live with the slower mechanism. DS 7.x does have the capability of using file cacheing - t...

Could you give us a sample input row and output row?

Disabling memory lookups will have a great performance impact, as you have seen. The "lock for updates" and disabling loading to memory will ensure consistency when more than one job is modifying and using the hash file concurrently - is this what is happening in your job? If you have only this job ...

Can you login on another developer's desktop using your login-id and compile the job? I think that permissions are the likely root of this problem, as the compilation doesn't access the desktop system.

Snassimr, there are several very different ways to do this that depend upon your data. Does the string already contain the delimiters or do you want to break it up into chunks of length n; is the string of fixed length? Have you looked into using the row splitter stage and why do you feel you can't ...