DSXchange

I think using below fuction i can achieve this change ereplace OR convert OR Left OR Right based on your requirement the User Define query should look like update tablexyz set key3=? where key1=? and key2=? and key3=? and this statement should be such that it can be parse by SQL engine, so any use ...

What are the keys for the update?


I will prefer to use Store procedure, as I get quite confuse with positional parameters and the way columns should be place in input stream for this type of update while using ODBC

What about the page break?

Play with the below function -

Here is another way to do - Consider your file contains Column Header. read the file seperate into two streams Stream1 will have column header (@inrownum =1) Stream2 will have data (@inrownum >1) for Stream1 - pivot the cloumn header custid key lname = lname mth = jan_sales, ...........,dec_pay assi...

Unless you specify a default for NullFieldValues within the Sequential file stage properties you will have trouble reading NULL's. I don't think so, if the field is set as Nullable = Yes, the NULL will be stored with nothing separated by the field delimiter (,) with the default setting something li...

Is it necessary to position the key fields at the first column grids when you are using ODBC stage for update?

As far as I know in MS OLEDB which also use the positional parameters, I had place the key fields at first of column grids for Update

karthi_gana wrote: 2) Get unmatching records

i think it may be a workaround for this problem...

That is what you are not getting in the job you have designed, so your work around will not work

I am sorry, if a order has multiple lines it is showing only one order line I think you are using the aggregate option in the transformation settings. Recently I faced the same issue, as the elements being grouped when there is a common value while using aggregate. And I have used a extra dummy ele...

That is so ugly. What's wrong with two jobs? Thanks Ray. We find that is the easiest way (in one job) to handle the overall DWH refresh as workaround; else creating two separate jobs and changing the job sequence of nearly 45 dimensions will take time to deliver, which we have to do eventually. Reg...

Thanks Craig

As per suggestion I will change the job design (not going for separate jobs)

INSSEQ file ----> XFM ------> MS OLEDB ------- dummy----->UPDSEQ file ----> XFM ------> MS OLEDB

Finally ... Our problem We have job design which inserts and update rows in tables reading data from sequential file as below INSSEQ file ----> XFM ------> MS OLEDB (insert rows in MS SQL DB Table) UPDSEQ file ----> XFM ------> MS OLEDB (Update rows in MS SQL DB Table) We get the following following...

chulett wrote:You mean, other than what has already been posted?

My problem is same - the jobs fails. Earlier I thought it was due to umlaut characters, now it is certain umlaut character is nothing to do with the failure

Now the job simply fails most often even without umlaut character in the input, run perfectly when I execute after failure with the same data.

The sequential file in the job is overwritten every night.
Error message is the same

Any idea how to resolve?

You say you don't think its related to NLS, yet the only time you have a failure is when there's an umlaut present? I'm not certain I follow your reasoning. Even if it is intermittent, it ONLY happens when there's an umlaut character present. Yes, because if run the same job/rerun whole refresh aga...