No, I hate this way of doing it.
Do this instead:
CREATE (me)-[r:has_story_2015_09_01]->(latest_story)
Then when you want the latest stories you can just query by the type and
date.
if you want a 2 days, you can use the "|" operator as an OR :
MATCH (me)-[:HAS_STORY_2015_09_01|HAS_STORY_2015_09_02]->(story)
RETURN story
ORDER BY story.published_date DESC
This way you are not deleting relationships all over the place which makes
a mess of data store.
You are using the relationship type as a sort of index.
On Monday, August 31, 2015 at 12:03:48 AM UTC-5, Vinicius Tavares wrote:
>
> Hello everyone! As I've read on Michael's pinned post, we can try to solve
> modelling issues here on this group.
>
> Here it goes:
>
> I'm trying to make a small Movies/TV Shows Social Network.
>
> The domain (simplified to focus on the problem):
>
> A *Profile* can *follow* other *Profiles* and also *Movies* and *TV Shows*
> .
>
> *Profiles* can *create* *Images* and *Articles* and *maybe* *mention*
> *Movies* and *TV Shows* on them.
>
> A *Profile* has a *News Feed* containing all *Media* and *Articles* that:
>
> - A* Profile* that he *follows* has *created.*
> - A *Movie* or *TV Show* he *follows* has been *mentioned* in.
>
> I had the idea that every media created (movie, article, image, video....)
> is part of a Profile's Story.
> So, when I post a new image mentioning an Interest (Movie, TVShow), I also
> create a Story linking them, as such:
>
> (Profile)-[has_story]->(Story)
> (Interest)->[has_mention]->(Story) // optional
> (Story)->[targets]->(Media)
>
> Stories will also be a linked list for both Profile and Mentioned Movies,
> for query speed purposes. So:
>
> (Movie)-[latest_mention]->(story)-[next]->(story2)-[next]->(story3)
>
> (Profile)-[latest_story]->(story)-[next]->(story2)-[next]->(story3)
>
> However, I'm struggling to UNION these two conditions. My current cypher
> is:
>
> START me=node(0)
>
> MATCH
> (me)-[:follows]->(interest)-[:latest_mention*0..1]->(story:Story)<-[:has_story]-(publisher:Profile)
> MATCH (story)-[:targets]->(media)
> RETURN story, media, interest, publisher
> UNION
> MATCH
> (me)-[:follows]->(publisher:Profile)-[:latest_story*0..1]->(story:Story)
> MATCH (story)-[:targets]->(media)
> OPTIONAL MATCH (story)<-[:has_mention]-(interest)
> RETURN story, media, interest, publisher
>
> ORDER BY story.created_time DESC
> LIMIT 15
>
> Is it a good approach? My write query seems to be quite ugly:
>
> START me=node({ me }), interest=node({ interest })
>
> OPTIONAL MATCH (me)-[r:latest_story]-(secondlateststory)
> DELETE r
>
> CREATE (me)-[:latest_status]->(latest_story:Story { story_payload })
> CREATE (me)-[:has_story]->(latest_story)
>
> WITH latest_story, collect(secondlateststory) AS seconds, interest
> FOREACH (x IN seconds | CREATE (latest_story)-[:next]->(x))
>
> WITH interest, latest_story
>
> OPTIONAL MATCH (interest)-[r1:latest_mention]-(secondlatestmention)
> DELETE r1
>
> CREATE (interest)-[:latest_mention]->(latest_story)
> CREATE (interest)-[:has_mention]->(latest_story)
>
> WITH latest_story, collect(secondlatestmention) AS seconds
>
> FOREACH (x IN seconds | CREATE (latest_story)-[:next]->(x))
> RETURN latest_story as new_story
>
> This query is adapted from the neo4j's cookbook about newsfeeds, merging
> with my need of mentions.
>
> Am I doing it right? Is there a shorter or more performant path to do so?
>
> Thank you for your help,
> VInicius
>
>
--
You received this message because you are subscribed to the Google Groups
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
