Issue 11913

Log messages for occurrence detail view show wrong taxon names

11913
Reporter: ahahn
Assignee: trobertson
Type: Bug
Summary: Log messages for occurrence detail view show wrong taxon names
Priority: Major
Resolution: WontFix
Status: Closed
Created: 2012-09-17 14:32:20.236
Updated: 2013-09-27 19:35:15.849
Resolved: 2013-09-27 19:35:15.805
        
Description: Log messages in a data usage log filter like http://data.gbif.org/datasets/resource/14186/logs/?event=3001-3999&logGroup=&logLevel=&sd_day=07&sd_month=08&sd_year=2012&ed_day=07&ed_month=08&ed_year=2012 show the wrong taxon names. The names shown in the record details (link from record id) are at the same time correct. However, the same log displayed the proper taxon names in August 2012 (see attachments; highlighted lines on second page of the pdf and in the png-file correspond with the following example on the database side). Similar effects in other datasets' logs suggest that something changed around late August/early September 2012 that caused a change in logged taxon ids, now pointing to wrong names for older log entries.

Example: occurrence_id 620785593
http://data.gbif.org/occurrences/620785593/
select id,data_resource_id,occurrence_id,taxon_concept_id,timestamp from gbif_log_message where occurrence_id = 620785593 limit 10;

id        | data_resource_id | occurrence_id | taxon_concept_id | timestamp
----------+------------------+---------------+------------------+--------------------
541599017 |            14186 |     620785593 |         60877755 | 2012-07-10 17:48:37
554842806 |            14186 |     620785593 |         60877755 | 2012-08-07 10:15:22
554843332 |            14186 |     620785593 |         60877755 | 2012-08-07 10:29:40
557031448 |            14186 |     620785593 |         60877755 | 2012-09-03 14:23:00
557031602 |            14186 |     620785593 |         60877755 | 2012-09-03 14:24:14
557436897 |            14186 |     620785593 |         60851864 | 2012-09-12 18:22:14
557490098 |            14186 |     620785593 |         60851864 | 2012-09-13 09:09:45
557730241 |            14186 |     620785593 |         60851864 | 2012-09-17 11:04:21
557732739 |            14186 |     620785593 |         60851864 | 2012-09-17 11:59:36
557732740 |            14186 |     620785593 |         60851864 | 2012-09-17 11:59:52

select tc.id as taxon_concept_id,tn.id as taxon_name_id,tn.canonical from taxon_name tn join taxon_concept tc on tn.id=tc.taxon_name_id where tc.id in (60877755,60851864);

taxon_concept_id | taxon_name_id | canonical
-----------------+---------------+------------------------------------------
        60851864 |        849623 | Aploparaksis brachyphallos (Krabbe, 1869)
        60877755 |       2716800 | Colella thomsoni

--> until Sept 03, 2012, the taxon concept id that was logged now links to "Colella thomsoni". From Sep 12, 2012, a different concept id was logged, which currently links to "Aploparaksis brachyphallos (Krabbe, 1869)"

--> however, a screenshot taken after 7th of August 2012 displays name for this occurrence as "Aploparaksis brachyphallos (Krabbe, 1869)", suggesting that the concept id that was logged before September 2012 used to link to this name, but now points to a different one.

The same is true for other records, and not only within this specific dataset: check, e.g., the links for the occurrence details via the record id vs the displayed names at log page http://data.gbif.org/datasets/resource/646/logs/?event=3006&logGroup=&logLevel=&sd_day=17&sd_month=01&sd_year=2011&ed_day=17&ed_month=01&ed_year=2012 (Santa Barbara) or http://data.gbif.org/datasets/resource/1083/logs/?event=3006&logGroup=&logLevel=&sd_day=17&sd_month=01&sd_year=2011&ed_day=17&ed_month=01&ed_year=2012 (Algaterra Movies). The change in taxon_concept_id is between August and September 2012.

reported by: Pascal Tschudin, GBIF-Switzerland


]]>
    

Attachment PortalLogsSep17.png

Attachment PortalLogsAroundAugust2012.pdf


Author: ahahn@gbif.org
Created: 2012-09-17 16:11:57.706
Updated: 2012-09-17 16:12:19.574
        
Checked the log merging process to exclude errors in the last merge/rollover, but there are no changes in the workflow. If the wrong source had been selected for the export, the concerned log records should come from about 8 weeks pre-rollover, but not further back than that. This is not the case, e.g. for AlgaTerra Movies: select id,data_resource_id,occurrence_id,taxon_concept_id,timestamp from gbif_log_message where occurrence_id =143162152;

    


Author: trobertson@gbif.org
Comment: no more table in new portal
Created: 2013-09-27 19:35:15.835
Updated: 2013-09-27 19:35:15.835