Can anyone help me write an xpath expression to select repeating nodes with duplicate cdata elements when they’re in staggered order

Question

<records>
    <record>
        <field_values>
            <field_value name="PSH" reportable="False">
                <item>
                <column name="procedure"><![CDATA[Duplicate1]]></column>
                <column name="comment"><![CDATA[]]></column>
                <column name="date"><![CDATA[1982]]></column>
                </item>
                <item>
                <column name="procedure"><![CDATA[Duplicate2]]></column>
                <column name="comment"><![CDATA[]]></column>
                <column name="date"><![CDATA[1999]]></column>
                </item>
                <item>
                <column name="procedure"><![CDATA[Duplicate1]]></column>
                <column name="comment"><![CDATA[]]></column>
                <column name="date"><![CDATA[1982]]></column>
                </item>
                <item>
                <column name="procedure"><![CDATA[Duplicate2]]></column>
                <column name="comment"><![CDATA[]]></column>
                <column name="date"><![CDATA[1999]]></column>
                </item>
            </field_value>
        </field_values>
    </record>
</records>

I am new to xpath expressions and I need to figure out an expression that selects duplicate items of this particular node to delete. I am only able to create the following expression:

/records/record/field_values/field_value[@name="PSH"]/item

Can someone assist me in finding a expression to select the duplicated items of the node?

Answer 1

distinct-values(/records/record/field_values/field_value[@name="PSH"]/item/column[@name='procedure'])