On 13 Aug 2008, at 14:43, Arek Kasprzyk wrote:
Hi Henrikki,
I think we understand your problem now. You have basically three
options
there:
1. Leave data where it is on all 7 instances and create a dataset
per each
one. Then deploy a portal connecting to those 7 instances and
presenting a
separate dataset for each. (you would have to run each source scheme
->
dataset conversion on each instance) - you can use Mbuilder to run
schema
transformation.
Hi Arek. Do I then understand correctly that there is currently no way
to have multiple remote (but identically-configured) marts appear as
*one* dataset* on a single portal mart? Basically like the existing
partitions I suppose, but with each partition residing in a remote
mart. Moving all data to one instance as Henrikki is basically forced
to do isn't always possible, e.g. for licensing reasons, or due to
large size.
Sorry if this has come up before on mart-dev and I'm repeating a
previous discussion, but are there any plans to support something like
this in a future version?
2. Move all data to one instance and and create a one transformation
file
with seven partitions.
(you need to choose one of the schemas and then provide a regular
expression
for detecting all your 7 source schemas. To try it out:
1. Add new schema and provide a regular expression so that Mbuilder
can
identify all 7 schemas)
You can run all transformation in as a single run. However this will
still
leave you with 7 different datasets.
3. Move all data to one instance and manually merge the data into
one mart
schema taking care of all the conflicting keys (Mbuilder will not
help you
with that because we do not support merges yet - it is in our to do
list)
I guess (3) is what you are already doing so we could not be of a
much help
to you this time. Sorry.
