Although the reason for asking the question is being addressed in a different thread, merge features with duplicate attributes in FGDB , I thought I would share there is a fairly simple way to select distinct, duplicate, triplicate, etc... records in a layer without having to create an additional table.
The following code snippet will select non-distinct records in a layer/table view based on comparison/selection fields.
def select_plicates_layer(layer, selection_fields, value=2, condition='ge'):
import operator as operator
from collections import defaultdict
if isinstance(selection_fields, basestring):
selection_fields = [selection_fields]
plicates = defaultdict(list)
with arcpy.da.SearchCursor(layer, ["OID@"] + selection_fields) as cursor:
for row in cursor:
plicates[row[1:]] += row[:1]
condition_oids = [oids
for oids
in plicates.values()
if getattr(operator, condition)(len(oids), value)]
if condition_oids:
oid_list = [oid for oids in condition_oids for oid in oids]
layer.setSelectionSet("NEW", oid_list)
return condition_oids
The real value of the function is less what it selects and the results it returns, i.e., the function returns a list of OID lists. Each OID list contains the OIDs of records that share the same values for the selection_fields. The list of lists can be used to determine how many duplicates, triplicates, quadruplicates, etc... are in the layer and what the makeup of them looks like.
The code is designed to operate on layers, so any existing selection of a layer is honored. The code operators more in a "SUBSET_SELECTION" manner than a "NEW_SELECTION" manner.
