XML parsing, return all elements except _children dynamically

Neo4j Graph Platform Import / Export
1

Hi

this code is
CALL apoc.load.xml ($mirror_filelocation, '')
YIELD value
UNWIND value._children AS plmxml
with plmxml
where plmxml._type = "ProductView"
with plmxml
return plmxml

returning a json like this. I want to return a map of all attributes except under _children dynamically without mentioning the name of attributes in my cypher.
i want to return
{
"primaryOccurrenceRef": "id10",
"_type": "ProductView",
"id": "id4",
"ruleRefs": "#id2",
"rootRefs": "id7"
}

how can i do it ?

the retuned json from above stmt is in below.

Thanks

{
"_children": [
{
"_type": "ApplicationRef",
"label": "xDVplLIJqd$DyB/QkqRC5OMJ6JDmB/AAAAAAAAAAAAAA/BOM/"
},
{
"_type": "UserData",
"_children": [
{
"_type": "UserValue",
"title": "BOM_precision_type",
"value": "imprecise"
}
],
"primaryOccurrenceRef": "id10",
"_type": "ProductView",
"id": "id4",
"ruleRefs": "#id2",
"rootRefs": "id7"
}

2

You can try using apoc.map.clean to remove the "_children" attribute, leaving all the remaining ones.

3

thank you Gary

apoc.map.clean is working but it is a function
i need to create a merge statment with remaining keys that are putput of apoc.map.clean.
how can i continue below to create a node at the end.

i have like this now
CALL apoc.load.xml ($mirror_filelocation, '')
YIELD value
UNWIND value._children AS plmxml
with plmxml
where plmxml._type = "ProductView"
with plmxml
return apoc.map.clean( plmxml , ["_children"], [""]) as output

output is returning this
{
"_type": "ProductView",
"id": "id4",
"ruleRefs": "#id2",
"primaryOccurrenceRef": "id10",
"rootRefs": "id7"
}

if i want to create a node with type that is value of output._type and also have attributes of the nodes exactly the same as keys mentioned in output, how i do it without hardcoding ?

Thanks a lot

4 
CALL apoc.load.xml ($mirror_filelocation, '')
YIELD value
UNWIND value._children AS plmxml
with plmxml
where plmxml._type = "ProductView"
with plmxml._type as label,  apoc.map.clean( plmxml , ["_children"," _type"], [""]) as properties
Return apoc.create.node([label], properties)

In this case you could replace '[label]' with '["ProductView"]' in this case, since plmxml._type always equals "ProductView". I left it generic incase you expand the 'where' condition to add extra types.

5

Thank you Gary, you are shaping the work nicely

assume executing your code, it creates many, nodes, all labeled : Occurrence. example of three nodes are in here.

the nodes have properties : id and occurrenceRefs. for example occurrenceRefs = "id93 id206"
those two ids in OccurenceRefs are child nodes of this node. it can be 0,1 or many childs mentioned in OccurenceRefs

for every node, i have two parse occurenceRefs to get those ids ( space separated, ) and then find associated nodes with those ids and create a CHILD relationship.

is there apocs to parse occurenceRefs to create a list and then how to do a for loop for all of those IDs in the list to do match and then create relationships?

match ( o:Occurrence) return o

{
"identity": 49,
"labels": [
"Occurrence"
],
"properties": {
"instanceRefs": "#id297",
"associatedAttachmentRefs": "#id78 #id86",
"occurrenceRefs": "id93 id206",
"_type": "Occurrence",
"instancedRef": "#id289",
"id": "id7",
"sourceRef": "#id288",
"parentRef": "#id42"
},
"elementId": "49"
}

{
"identity": 52,
"labels": [
"Occurrence"
],
"properties": {
"instanceRefs": "#id200",
"associatedAttachmentRefs": "#id97 #id104",
"occurrenceRefs": "id111 id151",
"_type": "Occurrence",
"instancedRef": "#id192",
"id": "id93",
"sourceRef": "#id191",
"parentRef": "#id7"
},
"elementId": "52"
}

{
"identity": 55,
"labels": [
"Occurrence"
],
"properties": {
"instanceRefs": "#id284",
"associatedAttachmentRefs": "#id210 #id217",
"occurrenceRefs": "id224 id239",
"_type": "Occurrence",
"instancedRef": "#id276",
"id": "id206",
"sourceRef": "#id275",
"parentRef": "#id7"
},
"elementId": "55"
}

6

while waiting for your help, i found i can split "id93 id206" to a list using
match ( o:Occurrence) return o, apoc.text.split( o.occurrenceRefs, " ") as splt

now i have two create relationship between the parent and two children in this list . somehow i have to iterate through elements in the list to repeat an action

it will be great you can help me.

thanks a lot

7

That is odd that you are getting 'Occurrence' nodes, when the condition is to filter through only 'ProductView' types. I assumed you changed that condition in your query.

Cypher has its own 'split' function. I used in the following query. The query extracts the id's, matches to find the corresponding Occurrence node, then creates a relationship between each Occurrence node and its parent node.

Let me know if it doesn't work.

CALL apoc.load.xml ($mirror_filelocation, '')
YIELD value
UNWIND value._children AS plmxml
with plmxml
where plmxml._type = "Occurrence"
with plmxml._type as label,  apoc.map.clean( plmxml , ["_children"," _type"], [""]) as properties
WITH apoc.create.node([label], properties) as node, split(properties['occurrenceRefs'], " ") as child_ids
UNWIND child_ids as id
match (o:Occurrence{id:id})
merge(node)-[r:HAS_CHILD]->(o)
set node.occurrenceRefs = null
return node, o, r