We have just encountered a batch of records in R6 where the observation comments made no sense.  They were all imported from three tab-separated text files in Sept 2009 using the import wizard.

In looking at the original text files it appears that in some cases (but not all) the comment from one record has been imported into the comment field of the following record.

There seem to be two failure modes:

• One is where a lengthy comment of one record has been truncated and re-used for the comment of the following record, replacing the record’s original comment. In this case the truncated comment is always 110 characters long and ends with three full-stops, i.e. "…". These can be easily found, e.g.

  SELECT COMMENT FROM TAXON_OCCURRENCE WHERE (([COMMENT]) Like "*..." And Len([COMMENT])="110");

  We have 86 such records, each being a truncated version of the prior record’s comment.

• The other mode is where the comments have "slipped" by one record, each acquiring the comment of the record before it. This is impossible to check for automatically. I don’t know how many of these we have.

Both failure modes have been seen within the same import, but neither is consistent. Some long comments haven’t been re-used in truncated form and some comments haven’t slipped onto the next record.

A re-importation today of one of the three text files (the largest) has gone in perfectly, so whatever the problem was (with the import wizard?) appears to have gone away.

I'm hoping the problem was restricted to just these three imports, which can easily be reimported using the current wizard.

I just thought I'd issue an alert in case this has affected anyone else unknowingly. It took us more than five years to notice!

Keith