I have, on occasions, encountered users who are contemplating importing Recorder data from spreadsheets using the Site ID and Record ID column types in the import wizard. The idea is to split the taxon occurrence key into Site ID and Record ID. This facility was provided to allow data to be imported from other databases using identifiers from the system concerned so that if corrections to existing observations are imported they will overwrite the original version rather than create new versions. It should NOT be used for data from another copy of Recorder (Recorder 2000/2002 or Recorder 6) because of the potential for corruption.

Site ID needs to be an 8-character NBN globally unique identifier for the system from which the observations originated and Record ID a 6-character identifier for the observation supplied by the source system. Record ID is only 6 characters because the additional 2 characters are reserved to allow unique 16 character keys to be created for entries in tables that may require multiple entries for a taxon occurrence, e.g. Taxon_occurrence_data, Taxon_occurrence_relation and Taxon_occurrence_sources, as explained in the Help (Search for Import Wizard and select the data format topic). The Site ID and Record ID columns in the import wizard should ONLY be used for data coming from systems other than Recorder and MapMate. Please note the warning in the Help regarding the need for Site IDs to be unique.

The potential for corruption if Recorder data imported via the wizard using Site ID and Record ID ever made its way into other systems that already contain data from the specified Site ID doesn’t bear thinking about. If you import a record with say, Site ID SR000247 and Record ID 004444, the survey event, sample and taxon occurrence will all have keys of SR00024700004444, unless there is already a suitable survey event and sample. The problem is that the survey event SR00024700004444 may exist in the original system but will probably be linked to different observations due to the way keys are created when data is entered via the standard means. Likewise with the sample. Hence, if this data was imported into a system containing the survey event/sample SR00024700004444 created by standard Recorder data entry one version would overwrite the other probably leaving taxon occurrences linked to the wrong samples.

The correct way to handle this situation is to go back to the holder of the data and get them to do a standard Recorder export in NBN Data (.xml) or, preferably, NBN Access Database (zipped) format and import that. If you don’t know who the owner is JNCC will be able to tell you who it is, provided the Site ID has been registered. If it hasn’t, they will, at least, be able to tell you who the reseller was, and they should be able to tell you who the owner is, subject to data protection issues.