Truncated node names in relational views can result in duplicate column names

[chiatt]

Relational view column names are created from slugified node names. If the node name exceeds the postgres column name limit, postgres simply truncates the name. However, this can result in two columns having the same name. For example:

If chemical_material_elements_found_data_assignment_statement_type_metatype gets truncated to:
chemical_material_elements_found_data_assignment_statement_type and a chemical_material_elements_found_data_assignment_statement_type node already exists in the same card, then the creation of the relational view fails.

To address this we need to check for duplicate nodes and make the duplicate name unique in some way to prevent collisions.

One way to do this would be to use aliases, but we would need to truncate aliases to 63 characters.

Another option would be to limit the length of node names to 63 characters. If more information is needed to describe the node than a 63 character title, it could be saved to the node description.

[chiatt]

Until this issue is resolved, database admins can run the following sql to identify which node names result in duplicates:

select * from (
	select nodegroupid, left(__arches_slugify(name), 63) as column_name, count(left(__arches_slugify(name), 63)) as duplicate_name_count, string_agg(name, ', ') as node_names from nodes 
	where length(__arches_slugify(name)) > 62
	group by left(__arches_slugify(name), 63), nodegroupid
	order by nodegroupid, left(__arches_slugify(name), 63)
) where duplicate_name_count > 1;