I have recently upgraded the Dragonfly Recording Network database from R2K to Recorder 6 by exporting from R2K to an NBN Access database (zipped) and importing into Recorder 6.  Each Vice-county was initially exported and imported back into an empty R2K database to identify errors which were then corrected before exporting and importing into Recorder 6.

The import wizard was used to import records from an Excel spreadsheet which had been created from a Biobase database.  All of  these were new records.  During the import the name of the survey was selected, names were matched with several new names created and unmatched locations were placed into the location name.  One invalid item was found which was an out of range observation date.  The report was as follows with no duplicate records:- 

22324 Items successfully read, of which there are:
    22324 New items.
    0 Duplicated items.
1 Invalid item.

The import had to be cancelled so that the date error could be corrected in the source spreadsheet.  The import wizard was run for a second time with the corrected date and produced the following analysis:-

22321 Items successfully read, of which there are:
    19135 New items.
    3186 Duplicated items.
0 Invalid items.

An example key was selected and found to belong to a record in a totally different survey.  The previous backup was restored and the records were imported successfully.

The above error was replicated with other datasets but some did not fail on the first cancel but failed on the second, third or fourth attempt.  A subsequent import of 1507 new records failed on the first attempt with 2031 duplicate items.  If the ‘Accept all original’ option was selected the records were imported without any measurement data.  It was in this instance impossible to apply the workaround described above.

Can anyone let me know what is happening and how I can in future reliably import new records.

Steve Prentice
Dragonflies in Focus Officer
Britsh Dragonfly Society