Walks with multiple samples for each section

[JimBacon]

A record should consist of a parent sample for a walk with one sub-sample for each section in that walk.

When records are imported and the occurrences for a section are not grouped together then the importer will create multiple samples for each section.

The following query finds 707361 instances of sections having multiple samples:

select parent_sample_id, location_id, count(id)
from cache_samples_functional
where survey_id = 64 -- UKBMS transects
	and parent_sample_id is not null
group by parent_sample_id, location_id
having count(*) > 1
order by count(*) desc

Details of how a similar issue was resolved at BiologicalRecordsCentre/ABLE#459