The "indexing" page of each dataset (e.g.
http://data.gbif.org/datasets/resource/1523/indexing/) allows access to the
event logs recorded during indexing. Linked from there, the "Process ID" in
the table leads to an overview of logged issues. Clicking on "view" from
there opens a full list of log messages for the selected error category.
This list contains record IDs that link through to individual records.
For collection curators working offline, a download option is provided:
"Download these logs" in the grey box on that error list. The download,
however, always contains empty values for a number of fields:
Scientific name, Occurrence record id, collection code, institution code,
catalogue number (sample file attached). Historically, this is because some
of the log messages are summary messages, giving only one message where it
is identical for a whole series of records. However, in all cases where the
"Count" is 1 (i.e., the message relates to exactly one record), all these
values should be included in the download. Only when this is the case does
the download have any value for a curator trying to correct the reported
issues in their database, identifying the concerned records quickly
Original issue reported on code.google.com by ah...@gbif.org on 7 May 2010 at 1:00
Original issue reported on code.google.com by
ah...@gbif.orgon 7 May 2010 at 1:00Attachments: