DSXchange

Hi, we have a job which is doing a lookup to 2 lookupfiles in one stage. one of the lookups will always have values , other only sometimes. Now: some of the fields in the lookup-files have the same name. input: field A,B,C lookup 1 : A(key),D,E (available for all records) lookup 2 : B(Key),D,F outpu...

For me the major difference is more in the configuration: in a parallel job you define the number of parallel processes at runtime (by giving the config file.) In server you define this when designing the jobs. So when you move from a 2 CPU machine to a 4 cPU-machine, most of the work for parallel i...

we are succesfully using oracle 10.2 with datastage 7.5.1A.
I just recieved an email from IBM about ds7.5.2 , which would have oracel 10.2 support. Maybe something to look at.

well, luckily for me this was a one-time only experience.
I didn't get the parallel read working on IQ. If you define more then one node, it just executes the full query on every node.

if you place an order stage part of this logic can be done in there. I'm not sure about the exact name but there is an option for a key-change column. This will be set to 1 if the previous record had an other key then this one.

something we did (but with lot less fields): in a transformer create the record with all the fields, so 96 times logic like : if time=001500 then amount else NULL for field amount0015,if time=003000 then amount else NULL for field amount0030,... after this do an aggregation where you summarize all t...

Hi, I'm working on data from multiple sources.(oracle/csv,..). I do some aggregations/transforms and load into an oracle db. In the DB one of the fields is defined as number(14,6). Throughout the DS-job the data for this field is always a double, without any further specifiactions. This seems to rou...

most common suggestion on this forum: split the inserts and updates in your job and do them seperatly.

Offcource always important: check your indexes(mostly for the updates)

I mean one lookup stage that does both lookups.

I cannot read the premium content, so sorry if this is a repeat.

You can do this in one lookup followed by a transform. Just do the both lookups to the file, so once on 2 keys, once on one key and a default. In the following transformer you put in logic: if lookup1 isnull then lookup2 else lookup1.

The most important reason for us to limit the number of projects are commonly used routines. If you write a routine that is used in multiple projects you have to copy it to all projects, which can become problematic if you have 50 projects.

Kumar_s, I'm not doubting the use of change capture, BUT: easy setup using change capture would be: extract target(before), extract source(after), go trouch change capture, apply changes. From my oracle experience(so again not sure if relevant on sql-server) even if you have exactly the changes it i...

A completely other scenario: If the number of inserts/updates is large and it's difficult to find the changed records: Just truncate and reload.
I'm not used to sql server but on an oracle this is the fastest solution once you reach about 10% inserts/updates.

c_wf, from experience:this is not as powerfull as it looks. It is quick and dumb partitioning. Example: if you extract from one table you mention the same table as partition table. Datastage will look for min and max rowid, split this up in even parts and then fire 4 queries with the added clause ro...

hi,
in your error message you see :

connect failed for env: null

Under your connection in the Oracle stage you need to set the remote server property to your database name. For now it has no clue in which DB to look.