dataware house typical file formats and extraction api? vs java capabilities

[NOBODY]

Hi,

In each of the following questions, I'm looking not only for things that
exists but things that are preferred by data warehouse users.


1-what is the usual approach to download data from a source like a data
warehouse?
-zip?
-tgz?
-gz?
-other compression/tar/jar/lha?

2-same question as above but when the source is known to be a java service?


3-what medium is preferred
-http request to a frontend (in my case a java exporting servlet)?
-ftp/tftp?
-scp?
-nfs/samba?

4-same question as above but when the source is known to be a java service?


5-what kind of export parameterization and conveniences do they offer?
-dumb flat listing and pull?
-time ranges and basic selection?
-extensive querying?
-stateful sessions, for example with cursors to continue where the
downloader left?


Thanks.

 

[Tajonis]

First of all a Data Warehouse is a specialized database for storing
large,large amounts of data. These database are built in a specialized
schema (i.e. star, snowflake, etc) that is designed for reporting.

1-what is the usual approach to download data from a source like a data

warehouse?

A data warehouse is built around existing data found with in another
database so
"downloading" data from a "source" like a data warehouse should not be
done. You should look at extracting the data from its original source
and not the data warehouse. If you must go against the data warehouse
then look into a nice OLAP tool.

2-same question as above but when the source is known to be a java
service?
Retrieving data from a java service should be dependent on the need at
hand. If the the data that is intended for "download" is not text the I
would recommend a TAR or ZIP format.

3-what medium is preferred
?What
You users should determine the medium

4-same question as above but when the source is known to be a java
service?
Same answer as above.

5-what kind of export parameterization and conveniences do they offer?
Same answer as above.

 

[NOBODY]

Reformulated:

What is the usual approach to gather data from sources?
I'm looking for what techniques typical data warehouses like to use for
pulling data.


What medium is preferred (http, scp, ftp, etc...)

What kind of export parameterization and conveniences do they expect?

Thanks.