DSXchange

Just as a question, what is the data source in this process? If its a DB there might be ways of avoiding a split fork join by incorporating the identification of potential duplicates into your select query.

Hi All I am having a little bit of an issue when executing a process to update a target MSSQL DB using an ODBC connector stage with the Write mode set to Insert New Rows Only. The issue is, that it works for some target tables and not others. We have checked for differences in the table set up betwe...

Thanks for the input Mike I was thinking about an explicit cast during selection, via a view. I was trying to avoid that as there would be 200+ tables for which I would have to write views, then there is the on going maintenance if they ever change the table structures (which they do constantly). I ...

Hi All I am facing an issue with implicit data type conversions and I am hoping that somebody can help. My job is a relatively simple one: ODBC(ora) ----> copy -----> ODBC(ora) The job is fully parameterised and RCP is enabled. The job itself has no columns specified. The job is to read data from on...

You could just add them to the parameter list and pass them as a parameter?

Or you could try GetEnvironment("NM_USER") - but not sure if that works in a sequence.

Does the routine only select the one value? Is it possible to pass out the value that is being passed to the log? Can you put an execute command stage in and just echo the parameter value as it passed to the job? Does any part of the sequence past the running of the routine run correctly? Is it alwa...

What value are you passing the parameter P_BATCHID and what data type is the parameter?

Have you already split the input columns into separate columns is it still a single column containing the comma delimited record? To do the looping: 1. Set up your 3 output columns: Name, Description, Phone 2. Set up 2 loop variables eg lvDescription, lvPhoneNum; Set the loop while to @ITERATION <=3...

Use a nested select statement where you add a row number:

select * from
(select key, row_number() over(partition by 1) as rownum from table1)
where rownum > 1;

If you set up your schema that way, then yes it will split it into 2 separate columns. But how do you determine what data is in which column? You will need the column headers to determine which input columns goes to which output column. Actually if you have column headers you can create a single job...

Sorry I wasn't clear on that.

For each input file that you receive, will it have column headers?

Do you have static column headers provided with the file?

From soure i have 4 columns and runtime i need to pick 1 key column and any 2 of those 3 columns. file is comma separated only,Please suggest me how to do this. When you say Any 2 of those 3 columns, are you suggesting that the input file could have a mix of columns? eg Row 1 of the file could be K...

Hmm. 2 ways spring to mind. Option 1 1. Sort by Column 2, then by column 1. Add a keyChange columns 2. In a transformer, on the input link re-sort by column 2, keychange asc. 3. In a stage variable (sv1), hold the value of column 2 where keychange = 0 4. In another stage variable have the logic: If ...

So the source is a DB? If that is the case - are you able to ensure that the first row selected is the header row? If you can you can just as easily exclude it in your select statement.