solgenomics / sgn

The code behind the Sol Genomics Network, Cassavabase and other Breedbase websites
https://solgenomics.net
MIT License
66 stars 35 forks source link

trial deletion from the back end doesn t operate on some trials #775

Closed bauchetg closed 3 years ago

bauchetg commented 7 years ago

Requested by NACCCRI: Using deletion_trial.pl script several trials do no get deleted, here is an example:

perl bin/delete_trials.pl -i 1608 -H db5.sgn.cornell.edu -D cxgn_cassava

Delete project entry... Cannot delete trial with associated phenotypes. Trial CBSD_GxE_Pallisa_2015A successfully deleted Retrieving trial information for trial 1608... CBSD_GxE_Serere_2015A, CBSD_GxE_Serere_2015A. Delete? Y Deleting trial CBSD_GxE_Serere_2015A Delete metadata... Use of uninitialized value $file_id in concatenation (.) or string at /usr/local/share/website/sgn/lib/CXGN/Trial.pm line 891, <> line 7. trying to delete association for file with id ... Deleting md_files linking table entries... Deleting phenotypes... GOING TO REMOVE 0 EXPERIMENTS... Deleting layout... _delete_field_layout_experiment... Attempt to delete field layout that still has associated phenotype data. ERROR {UNKNOWN}: cannot delete because of associated phenotypes at /usr/local/share/website/sgn/lib/CXGN/Trial.pm line 777

bauchetg commented 7 years ago

After deleting: DELETING 1 phenotypes Deleting linking table entries... RETRIEVED 0 METADATA ENTRIES... Use of uninitialized value within %projects_that_are_crosses in numeric eq (==) at lib/CXGN/BreedersToolbox/Projects.pm line 162.

bauchetg commented 7 years ago

Seems that some previous data was loaded using trait code instead of variable code, ex: Plant Stands Harvested Counting was coded CO:0000434 when it expects CO:0000010 Similar problem as for NRCRI datasets no?