1

Topic: A very odd story

I noticed on  my download log today that 2 records of Rosaceae had been taken from the HBRG Insects dataset.  That got me wondering, as you might imagine.

The result of my research is this:

On 27 Nov 2008 I imported a batch of records to R2002 via the Import Wizard, including two records of the moth Coleophora lusciniaepennella.  I know these were correctly stored, as I have an export taken immediately afterwards (I do this for every import to R6 for checking and backup purposes).

On seeing the surprising news that Rosaceae had been reclassified as insects, I have established that the Coleophora lusciniaepennella records had morphed into Cotoneaster rotundifolius.

Since being entered, the data were imported from R2002 to R6 (this happened in Dec 2008 or Jan 2009), and R6 and the Dictionaries have been updated as updates became available.

Looking at Gateway 4, it is apparent that the corruption occurred before 26 July 2013, but without a major investigation of backups I can’t get any closer.

I have now corrected the dets in R6 and made a note in the dataset metadata (the sooner we get back the ability to make comments on Gateway records the better), so I don’t feel I need any more.

However, if it happened to me, it could happen to anyone else, I imagine, so I am documenting this for reference and for the information of the Recorder team.

Murdo

2

Re: A very odd story

I have had a look at this and found that you are inded correct. Back at dictionary J Taxon_List_Item key NHMSYS0100001982 with a  TV key of NBNSYS0100002527 was Coleophora lusciniaepennella in the Lepidoptera Micro list. Now  Taxon_LIst_Item key NHMSYS0100001982  with a TV key of NHMSYS0100001124 is Cotoneaster rotundifolius on  the NBN Gateway Vascular Plant Dataset. The change happended in Dictionary upgrade Z  with and entry dated 4/11/2011.

I have  compares the Dictionary as at J with the changes made in update Z and this is the only change to species which looks odd. However,  Taraxacum at Dictionary J has become Artrolobium in Z. Doubt if this will have any impact on data input, but not  sure what effect it had on the list. I would think this perhasp arises froma manual cahnge made by the NHM where keys got muddled up.

Perhaps we should give some though to checking that changes  made to the TV keys for a TLi key in an update are reasonable.

Mike Weideli

3

Re: A very odd story

Thanks, Mike.  It seemed such a bizarre event that I was wondering about my sanity.  Good to have a rational explanation.

M.

4

Re: A very odd story

I have run a complete check from Dictionary J to the latest version looking for anything similar and fortunately I can't see anything. Changing the TV key for TLi key isn't something done very often and there were only 484 changes from J. Of this 294 had kept the same taxon name, which left only a few hundred to look through and these all seem reasonable changes apart from the one you found.  I will add a check into the dictionary processing to makle sure it doesn't happen again.

Mike Weideli

5

Re: A very odd story

That's a relief.  There seem only to be two other data holders for C. rot., one of them BSBI who we may assume will not have any moth records, so not a major problem.

M.

6

Re: A very odd story

To round this one off, problem solved thanks to Mike.

M.