taupilu / gbif-providertoolkit

Automatically exported from code.google.com/p/gbif-providertoolkit
0 stars 1 forks source link

Ability to delete mapping or re-execute automapping #1020

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
Hi,

Here at BeBIF, we extensively use the automapping features for DarwinCore terms 
to make long-term support. Concretely, we create a (sometimes complex) SQL view 
that map our database to DwC terms expected by IPT. The IPT configuration is 
therefore straightforward: when we create the mapping, all fields are 
auto-detected 

But there are cases where the fact that mappings cannot be removed or simply 
auto-remapped (like on mapping creation) is a bit difficult. I for example have 
a dataset with rich metadata, but I added a few more DwC fields to the view. If 
I want to test that my view still fit the automapping, I have to destroy and 
recreate the whole resource, which mean I'll have to re-enter the metadata.

When investigating this, I also discover that (after completion of the 
auto-mapping), the mapping use db column position in the mapping (and not 
column name). So if I add a new field at the "beginning" of my view, all fields 
get shifted and the mapping is totally incorrect.

Alternatively, for my specific case here, I'd be interested in a procedure to 
manually remove an existing mapping by hacking into the IPT data dir (bad, I 
know...)

thanks a lot

Original issue reported on code.google.com by nico...@gmail.com on 26 Nov 2013 at 4:16

GoogleCodeExporter commented 9 years ago
Thanks for taking the time to report your issue. It's a good suggestion, and 
will be under consideration. Given that you have some experience in interface 
design yourself, and are very familiar with the IPT, I welcome detailed ideas 
of how the interface could be adapted to adopt this new feature.

Original comment by kyle.br...@gmail.com on 26 Nov 2013 at 4:43

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
Thanks Kyle !

For the interface, I won't go for anything complex here: either a "Delete" next 
to the Edit button (right) in the mapping section of the resource overview, 
either an "re-do automapping" one on the mapping page. In both cases, some 
"This is a destructive action, warning, blablabla" confirmation dialog should 
pop up.

During this small experience, I also noticed IPT let us configure two different 
Core mappings for a given resource... It may be a bit more solid and 
idiot-proof if it it just prevent that. Do you want to open a different ticket 
for this suggestion ? 

Original comment by nico...@gmail.com on 28 Nov 2013 at 10:22

GoogleCodeExporter commented 9 years ago
You don t have to destroy and recreate the whole metadata. I just delete the 
mapping and re-analyse the source. This usually works for me, i never had to 
retype any metadata...

Original comment by dimibros...@gmail.com on 28 Nov 2013 at 10:24

GoogleCodeExporter commented 9 years ago
@Nicolas, thanks for the ideas. Simple is best I agree. 

As for the multiple core mappings, it's allowed because there are cases when 
publishers have to break a large source file into chunks. In such cases, they 
will have multiple core mappings that serve to concatenate those chunks 
together. 

Original comment by kyle.br...@gmail.com on 28 Nov 2013 at 11:46

GoogleCodeExporter commented 9 years ago
Dimi is of course right, the delete mapping button already exists, I only 
didn't noticed it. So this ticket is actually invalid. I do however think the 
this button can be better placed in terms of user interface.

It's currently located in the higher part of the mapping page, under the Filter 
field. It would be better at the bottom of this page, or even better on the 
resource overview page (next to the edit mapping button). 

What do you think ?

Original comment by nico...@gmail.com on 29 Nov 2013 at 1:38

GoogleCodeExporter commented 9 years ago
Thanks. Marking as invalid therefore. 

If further improvement can be made to the interface, let's create a separate 
new issue. Best wishes

Original comment by kyle.br...@gmail.com on 3 Feb 2014 at 5:36