06-07-05 07:47 AM
(I originally posted this message in the Google BizTalk Group, but was
hoping for some wider exposure; I apologize in advance for the
crossposting.)
My company is looking to enhance our transformation services platform,
and we're considering moving to a BizTalk solution. We have no
in-house BizTalk expertise, so we're really just now looking around.
Our current transformation scenario involves transforming large chunks
of SQL data to smaller standardized chunks of data using DTS packages.
One of the chanllenges that we face is the fact that our source data is
a syslog stream, which depending on the contents may actually represent
any one of several different entities. We also want to group like
events by the minute, so that if we have 5 similar events in a minute,
we get on record in the database that shouw 5 occurreces.
How we do this now is using SQL Server DTS to:
1. Bulk Load the data into a staging table; during this bulk load, the
syslog field is parsed into it's various elements using VBScript
ActiveX components.
2. Select from the staging table and grouping the data for it's final
insert into the target table.
I have no clue where to begin with BizTalk 2004 to approach this
functionality; I have played around with examples, and figured out to
build an orchestration using a map to transform a dataset from one SQL
Port to another; however, the message that arrives on the inital port
contains similarly structured events. In our case, there is no
guarantee that the data will be similarly structured. Rather, the
initial message is more like a container of smaller messages, each of
which must be interpreted and mapped to a new message. Is there q way
to do this with Biztalk?
I realize this is a convaluted explanation, but please have pity on me.
I'm just now diving in to BizTalk, and I am drowning.
TIA,
Stu
[ Post a follow-up to this message ]
|