Re: [U2] UPLOAD not DOWNLOAD

2012-10-27 Thread Wjhonson
I did mention in my initial posting Cederville's DOWNLOAD
Which extracts data *from* Universe.
It would be counter-intuitive to call a system which loads data *into* Universe 
also a Download.
Rather it must be called an Upload.


 

 

 

-Original Message-
From: Tony Gravagno <3xk547...@sneakemail.com>
To: u2-users 
Sent: Sat, Oct 27, 2012 12:23 pm
Subject: Re: [U2] UPLOAD not DOWNLOAD


Well, what you say below is exactly opposite of what I described.
Normally I would encourage anyone else to re-read what I wrote and
cite specific text which leads to the conclusions below - in this
case... nah.

The data source never needs to know anything about the target - that
would be silly.

And no, I didn't describe a process that "goes the wrong way" - you
said "*into* Universe (not out from it)" which is what I described.
Perhaps what confused you is that the code for going in both
directions looks exactly the same - which is one of the neat aspects
of the approach.

BTW, someone might suggest from my code that the retrieval from the
source is done in bulk while posting  to the U2 target seems to be
done with individual transactions. That could be the case and for
large loads I'd modify the code to aggregate the data into a
collection and then post that to U2 in a single transaction.

BTW, the terms UPload and DOWNload are hugely ambiguous, and I believe
you have defined the terms backwards here. When I download to my PC,
it is TO my PC, not away from it. When I upload to a web server, the
data is moving out. And yet you define "UPLOAD not DOWNLOAD" as
meaning  "*into* Universe (not out from it)". The terms IMport and
EXport are more explicit. Your reference to the Cedarville Download
utility is a clue to the confusion. Cedarville is using Download in
the context of the recipient of the information downloading
information to a more familiar format from the Unidata server. From
the perspective of everyone in this group, that would be an Export,
and going in the opposite direction would be an Import. So what you're
describing is an Import into U2 ... traditionally described as a
download.

T

> From: Wjhonson
> 
> That goes the wrong way.
> Download does not depend on knowing who is consuming what to
> create it's results.
> 
> So Upload should not need to have interactive consumption details
> either.
> That's my take.
> 
> It's reads a source and decides on its own without the sources
> assistance where to put what.
> 
> You're coming at it more from a point of view of both sides
*agreeing*
> on what data is in what position.

___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users

 
___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users


Re: [U2] UPLOAD not DOWNLOAD

2012-10-27 Thread Tony Gravagno
Well, what you say below is exactly opposite of what I described.
Normally I would encourage anyone else to re-read what I wrote and
cite specific text which leads to the conclusions below - in this
case... nah.

The data source never needs to know anything about the target - that
would be silly.

And no, I didn't describe a process that "goes the wrong way" - you
said "*into* Universe (not out from it)" which is what I described.
Perhaps what confused you is that the code for going in both
directions looks exactly the same - which is one of the neat aspects
of the approach.

BTW, someone might suggest from my code that the retrieval from the
source is done in bulk while posting  to the U2 target seems to be
done with individual transactions. That could be the case and for
large loads I'd modify the code to aggregate the data into a
collection and then post that to U2 in a single transaction.

BTW, the terms UPload and DOWNload are hugely ambiguous, and I believe
you have defined the terms backwards here. When I download to my PC,
it is TO my PC, not away from it. When I upload to a web server, the
data is moving out. And yet you define "UPLOAD not DOWNLOAD" as
meaning  "*into* Universe (not out from it)". The terms IMport and
EXport are more explicit. Your reference to the Cedarville Download
utility is a clue to the confusion. Cedarville is using Download in
the context of the recipient of the information downloading
information to a more familiar format from the Unidata server. From
the perspective of everyone in this group, that would be an Export,
and going in the opposite direction would be an Import. So what you're
describing is an Import into U2 ... traditionally described as a
download.

T

> From: Wjhonson
> 
> That goes the wrong way.
> Download does not depend on knowing who is consuming what to
> create it's results.
> 
> So Upload should not need to have interactive consumption details
> either.
> That's my take.
> 
> It's reads a source and decides on its own without the sources
> assistance where to put what.
> 
> You're coming at it more from a point of view of both sides
*agreeing*
> on what data is in what position.

___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users


Re: [U2] UPLOAD not DOWNLOAD

2012-10-27 Thread Wjhonson
Did you read Brian Leachs document?
I think the way he solved this was rather clever, using a TCL command to 
associate the columns to attributes, then once you've done it once per 
definition, you can set it and forget it.

Much more clever than writing an import routine for every new type of document.

 

 

 

-Original Message-
From: George Gallen 
To: U2 Users List 
Sent: Sat, Oct 27, 2012 6:52 am
Subject: Re: [U2] UPLOAD not DOWNLOAD


Will, if you don't want interactive (manual mapping), there is no other way 
aside from using DICT items that can be used to automatically map from an input 
file to your UV file, unless you feel like re-programming and compiling each 
time to change which columns go to which attributes. And you still may wind up 
having to manually map anyway, if your DICT name/description don't match column 
headers at all. 

I'm sure you have converted your share of client supplied files, and they are 
rarely the same, in column heading names or order of columns, even the 
formatting of dates, countries, states, phone numbers.

Outputting is easier, it's usually, here's my data, now you deal with it to 
import it

George Gallen
Senior Programmer/Analyst
Accounting/Data Division
ggal...@wyanokegroup.com
ph:856.848.9005 Ext 220
The Wyanoke Group
http://www.wyanokegroup.com

From: u2-users-boun...@listserver.u2ug.org 
[u2-users-boun...@listserver.u2ug.org] 
On Behalf Of Wjhonson [wjhon...@aol.com]
Sent: Friday, October 26, 2012 11:40 PM
To: u2-users@listserver.u2ug.org
Subject: Re: [U2] UPLOAD not DOWNLOAD

That goes the wrong way.
Download does not depend on knowing who is consuming what to create it's 
results.

So Upload should not need to have interactive consumption details either.
That's my take.

It's reads a source and decides on its own without the sources assistance where 
to put what.

You're coming at it more from a point of view of both sides *agreeing* on what 
data is in what position.

___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users
___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users

 
___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users


Re: [U2] UPLOAD not DOWNLOAD

2012-10-27 Thread George Gallen
Will, if you don't want interactive (manual mapping), there is no other way 
aside from using DICT items that can be used to automatically map from an input 
file to your UV file, unless you feel like re-programming and compiling each 
time to change which columns go to which attributes. And you still may wind up 
having to manually map anyway, if your DICT name/description don't match column 
headers at all. 

I'm sure you have converted your share of client supplied files, and they are 
rarely the same, in column heading names or order of columns, even the 
formatting of dates, countries, states, phone numbers.

Outputting is easier, it's usually, here's my data, now you deal with it to 
import it

George Gallen
Senior Programmer/Analyst
Accounting/Data Division
ggal...@wyanokegroup.com
ph:856.848.9005 Ext 220
The Wyanoke Group
http://www.wyanokegroup.com

From: u2-users-boun...@listserver.u2ug.org 
[u2-users-boun...@listserver.u2ug.org] On Behalf Of Wjhonson [wjhon...@aol.com]
Sent: Friday, October 26, 2012 11:40 PM
To: u2-users@listserver.u2ug.org
Subject: Re: [U2] UPLOAD not DOWNLOAD

That goes the wrong way.
Download does not depend on knowing who is consuming what to create it's 
results.

So Upload should not need to have interactive consumption details either.
That's my take.

It's reads a source and decides on its own without the sources assistance where 
to put what.

You're coming at it more from a point of view of both sides *agreeing* on what 
data is in what position.

___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users
___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users


Re: [U2] UPLOAD not DOWNLOAD

2012-10-27 Thread Manu Fernandes
hi John,

Depending what you are looking for,
- a generic "ETL" tool where you define
1) Extract data from a defineable source (remote sql, csv-xls-dbf-xml-json 
files, webservice result , ...) with extendable method
2) Transform the received clob to internal UV data representation (iconv, mv'ed 
representation, build multi-part unique key, ...)
3) Load into U2 files (update one or more uvfiles into a Transaction (db 
transaction + application transaction))

That's not  a "routine" that's  complex tool with a lot of params and 
subroutines :-p

I have it, writen in SB+, require SB+ at runtime, but it's not free. Please let 
me know if you can be interested.

-

- Anyway, If you search routines to import, write one  record on one uvfile per 
"received record",  from some specific data source I'm ready to share my 
experience (sql, dbf, csv, mdb, xls, xml, ...).

Please put some detail of the source.

regards
Manu

Wjhonson  a écrit :


Does anyone have a robust ETL routine for loading data *into* Universe (not out 
from it).  Something like Cedarville's DOWNLOAD but in reverse ?

Before I roll my own.
___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users

___
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users