How to Join the multiple Database sources

[irajasekharhexa]

Hi All,

Can u any body explain how can we do the joins in the Server jobs if the data is to be extracted from multiple sources like Oracle, Sql Server and Sybase,flat files etc.,

Thanks in Advance.

Regds

[ray.wurlod]

In server jobs you use the Transformer stage with one stream input link and other reference input links.

The stream input link is delivered a stream of rows from one data source (some joins may have been performed there). Each reference input link connects to another data source and is provided with one or more key values (per row) with which it performs a "get row that has this key" function.

If the data sources are remote, or do not support key-based lookups (such as flat files), you would prefer making a local copy for the reference inputs. This is precisely the function of the Hashed File stage; hashed files provide the fastest lookup-by-key capability.

[rkdatastage]

Hi
here when u are going to work with hetrogenous data sources
there is no direct stage where u can connect and extract data. you have to manually design the logic and extract the data. i think this feature has to be inculded in coming versions of datastage.

RK

[irajasekharhexa]

In server jobs you use the Transformer stage with one stream input link and other reference input links.

The stream input link is delivered a stream of rows from one data source (some joins may have been performed there). Each reference input link connects to another data source and is provided with one or more key values (per row) with which it performs a "get row that has this key" function.

If the data sources are remote, or do not support key-based lookups (such as flat files), you would prefer making a local copy for the reference inputs. This is precisely the function of the Hashed File stage; hashed files provide the fastest lookup-by-key capability.

[irajasekharhexa]

Thank you verymuch.

In server jobs you use the Transformer stage with one stream input link and other reference input links.

The stream input link is delivered a stream of rows from one data source (some joins may have been performed there). Each reference input link connects to another data source and is provided with one or more key values (per row) with which it performs a "get row that has this key" function.

If the data sources are remote, or do not support key-based lookups (such as flat files), you would prefer making a local copy for the reference inputs. This is precisely the function of the Hashed File stage; hashed files provide the fastest lookup-by-key capability.

[vmcburney]

The IILive2005 had a session on using the full IBM Information Integration suite (not just the Ascential products). With Information Integrator you can create a join across heterogeneous data sources and then process this data with DataStage. You can also use Information Integrator replication to trickle feed a data warehouse.