DSXchange

Maybe the "Replace empty value with NULL" should read "Replace empty value with NUL"? Please continue to investigate, and share your results. If nothing else you will create the reproducible cases demanded by IBM support. Ok A bit more digging and testing and I have more informa...

Char(0) is not NULL, it is NUL. It is a valid character from Control Set 0. When reading/writing sequential files you specify how NULL is represented. Ok then - now for my next question. How did this NUL value end up in the data. It wasn't parsed in the XML (there was no value for this field) so th...

Hi All Weird things are a foot - or at least I think that they are weird. I have a job which reads an input XML file and extracts several Seq files. Subsequent jobs then process these sequential files. So far so good. However I have a situation where a record in the XML file does not come with a par...

kduke wrote:Cobber is a good thang, right?

A cobber is your friend. Although you are just as likely to call your mates a b@stard as a term of endearment in Aussie lingo

If your lookup is a hash file you can use a UV stage to bring back multiple rows (I think). Never actually done it myself, but I am sure I have seen many postings about it on the forum.

For First Insertion I am assigining the Flag to "Y"..(NEW) If user changes the current_ind in target table to "N" my job should be update only that record with flag set to "N" Yes I got this bit the first time. To update your target table though you will need the key f...

Hi I have some Insert and update statements in Oracle DB, which is taking too much of time which id not affordable. The table is having lacs of records. Do you have any option to optimise it. Thanks and Regards Shabnam You will probably have to provide a little more information than this before peo...

For First Insertion I am assigining the Flag to "Y"..(NEW) If user changes the current_ind in target table to "N" my job should be update only that record with flag set to "N" I only ask as the above assumes that you have the key to update an existing record. I just th...

I am having my job as follows: I/P -->TX(lookup + constraint in TX for INSERT AND UPDATE)-->(NEWKEY GENEARTION(for insert new rows) --> Target table My question is if Flag = "N", then it is an old record and needs to be updated. How to accomplish this? Right now I am having only two links...

Ok this is a little embarrasing, but its seems that I have been misleading everybody a little :oops: Turns out that the Array size is set to 1, not 1000. (Its true what they say 'When you assume you make an ASS out of U and ME) Pretty sure that this could have a detrimental effect. The reason that i...

chulett wrote:Keep in mind the fact that you aren't being asked to change it forever, just to see what affect it has on your job.

Without a doubt - just testing at present.

pradkumar wrote:Hi Carter

I tried outer join in merge, but was not sucesful.
I am using tables as the source. It is giving error as invalid file name

Pretty sure you have to use sequential files for the Merge stage.

Thanks for all your input guys.

It seems the general 'vibe' here is to try a transaction size. Initially it was set to 0 for ease of re-run but since we are checking to see if a record already exists and filtering it, that seems kind of pointless.

Will give it a go and let you know how it all went

Generaly the increasing number of stream should be an advantage and not a dissadvantage. Give me more details on your streams. Ok maybe i have misled a little. The job looks something like Seq1 ---> Trf1 ---> Trf2 ---> Trf3 --\ \ / \ \ / \ OraOCI8a OraOCI8b / \ / / \ / Seq2 ---> Trf4 ---> Trf5 --->...

Actually the same message was posted on one of the original threads.