[seek-dev] RE: Failed to update a document in put interface.

Bing Zhu bzhu at sdsc.edu
Wed Sep 29 11:11:30 PDT 2004


Jing,

Can you modify a Metacat object with the current implementation of 'put'
with the following
object id as ' jing.102.2'. If yes, this will be the way to modify a Metacat
object. And this
requires user knew the current version of that object in Metacat to be able
to modify an object.

Otherwise, we need to discuss it with Matt to see what's his original
design.

BTW, In SRB, the modification of a dataset can be handled in the way by
passing a
filename argument similar to cgi-request with 'override' parameter. e.g.
   To modify an SRB object 'dt.txt',
   obj_id =
srb://username:password@srb.sdsc.edu:6613/home/whywhere.seek/dt.txt?override
=true

Sincerely,
Bing



-----Original Message-----
From: Jing Tao [mailto:tao at nceas.ucsb.edu]
Sent: Tuesday, September 28, 2004 9:08 AM
To: Bing Zhu
Cc: metacat-dev at ecoinformatics.org; seek-dev at ecoinformatics.org
Subject: RE: Failed to update a document in put interface.


Hi, Bing:

I don't think we need add a new parameter in "put" operation. I am
thinking if we need handle this situation(just local id):
1) user inserted jing.102.1 (version 1 of jing.102)
2) user inserted jing.102.2 (version 2 of jing.102)

In other word, do we need handle version issue in objectId in "put"
operation? I think you guys talked about this issue in Scottland. But I
forgot what is the decision.

Thanks.

Jing

On Mon, 27 Sep 2004, Bing Zhu wrote:

> Date: Mon, 27 Sep 2004 18:49:48 -0700
> From: Bing Zhu <bzhu at sdsc.edu>
> To: Jing Tao <tao at nceas.ucsb.edu>
> Cc: metacat-dev at ecoinformatics.org
> Subject: RE: Failed to update a document in put interface.
>
> Jing,
>
> Thanks for the testing and feedback.
> For the update part, I am not sure if we need to implement 'update' in the
> 'put'
> or it should be a separate operation in Ecogrid put service. This raises a
> question
> If yes, how the information of 'update' action mode (same as 'delete'
> action) to
> be passed in our current 'put' operation design.
>
> Here is the original design for put operation with our Ecogrid put
service.
> put(object, objectId, type, sessionId) throws EcoGridFault,
> InvalidSessionFault
>
> The only solution I can see is passing following message in 'type'
>     dataUpload, dataUpdate, dataDelete,
>     metadataUpload, metadataUpdate, metadatDelete
>
> Rather than the type as we designed previously as either 'data' or
> 'metadata'.
>
> Let me know what you think.
>
> Bing
>
>
>
>
>
> -----Original Message-----
> From: Jing Tao [mailto:tao at nceas.ucsb.edu]
> Sent: Monday, September 27, 2004 6:10 PM
> To: bzhu at sdsc.edu
> Cc: metacat-dev at ecoinformatics.org
> Subject: Failed to update a document in put interface.
>
>
> Hi, Bing:
>
> After installed query, authentication and put interface in my local gt
> 3.2, I created a junit test for testing authentication, put and query
> operations in metacat ecogrid node.
>
> Those interface works fine except when I tried to update a metadata
> document or data file.
>
> For example,  I inserted a metadata document jing.78.1 successfully by
> calling put method. Then if I tried to update the document to jing.78.2 by
> calling put method, I will get an error:
>  <?xml version="1.0"?>
> <error>
> Accession number jing.78 is already in use
> </error>
> Did you get this kind of error in your testing?
>
> By the way, I slightly revised some methods in EcogridPutClient and
> EcogridAuthClient: such as adding public keyword to methods rather than
> without any keyword(if without keyword, you can only access those methods
> in package), making login_action will return a sessionid which we
> can reuse it... I already checked in those modification into cvs.
>
> Thanks.
>
> Jing
>
>
> --
> Jing Tao
> National Center for Ecological
> Analysis and Synthesis (NCEAS)
> 735 State St. Suite 204
> Santa Barbara, CA 93101
>

--
Jing Tao
National Center for Ecological
Analysis and Synthesis (NCEAS)
735 State St. Suite 204
Santa Barbara, CA 93101




More information about the Seek-dev mailing list