Hi,

On Thu, Aug 3, 2017 at 10:34 AM, Farid <[email protected]> wrote:

> Dear Michael,
>
> Thanks for the answer.
>
>
> 1) It's already deprecated, yet I can't find any documentation nor real
>>> examples using its successor: neo4j-admin import
>>>
>> neo4j-admin import has built in help, but it's mostly just a different
>> script around the neo4j-import code.
>>
>
> That's the complain, it's very much Linux like, where some documentation
> can be found on the help, but nothing on the website, and no tutorials.
> For people used to the tool and have been working with it for long, that
> documentation does make sense, but for people new to the technology, it's
> too abstract and confusing (and I have 20 years+ of dev experience).
>

I will send the feedback to our docs team.

What kind of docs would be more helpful? Do you have good examples for this
kind of tool?


> 2) Even current documentation is limited to a very simple 2 nodes types
>>> and a simple relation between them, can't find how to use a more complex
>>> graph.
>>>
>> here is some other documentation:
>> https://neo4j.com/developer/guide-import-csv/#_super_fast_ba
>> tch_importer_for_huge_datasets
>> https://neo4j.com/blog/import-10m-stack-overflow-questions/
>>
>
> Thanks for the links, that I've seen as well as some other use case.
> For my understanding, this still a simple relation between 2 different
> nodes. What I mean by more complex graph is something that can be done in
> Cypher, such as some "if/else" comparisons, etc.
>
> I'll give it another shot, even if it means duplicating the data or so
> (cases when multipl ids are present, for example, a "Content" can have
> "Creator ID", "Category ID", "Size ID", so being able to match with more
> than 2 nodes at a time would be awesome)
>

It is not limited to 2 nodes and 1 relationship, you can import many
different node-types and relationship-types at the same time,
Just a relationship is always between two node-ids, in your case you would
have one rel for content->creator, one for content->category one for
device->size (although I would store the size as property tbh).

There are no conditionals in the import tool.

>
> 3) Import is supported only on an empty database, so there is no way to
>>> use it to import a batch of data into existing database, or create a
>>> complex graph by importing nodes and more complex relations.
>>>
>> Yes, this is currently a limitation, but it is meant to be accomodated in
>> the future.
>>
>
> Thanks, that's a great news, hopefully it's a near future, as it's highly
> important when needing to work with high frequency data that needs to be
> imported in batches.
>

I hear you.

I still want to see how we can improve your cypher import performance, so
it would be good to get a realistic sample of data + your existing cypher
import scripts.

The other option for such a dedicated use-case would be to create a custom
importer which uses the Neo4j Java APIs that allow you full control over
concurrency, batching and queuing, like here:
https://maxdemarzi.com/2016/09/26/custom-importers/

>
> Sincerely,
> Farid
>

-- 
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.

Reply via email to