Type woes loading X3D JSON

Elastic Stack Logstash
1

I am experiencing type woes loading X3D JSON. I think because it's a very hierarchical document, and there's no schema for it yet. There is an XML schema, which might be used to load metadata into elastic search, but it might be easier to modify the source JSON. I was wondering if there was a way to report the type conflicts more descriptively? I'm getting very long errors that are hard to deal with.

Thanks,

John

2

It's not clear to me what you're asking. You have a JSON document that you want to validate against an XML schema? Or do you have particular changes you want done to the document before you pass it to Elasticsearch?

3

I’m trying to load 115 documents into ElasticSearch with LogStash and getting error 400, along with a lot of successes. I think it’s a typing issue. I am not the source of the document. I might be able to modify the document, bug first, I need to know which attributes are failing with typing issues. The errors are hidden in the JSON error output I think, but I’m not sure.

John

4

The Logstash log should contain clues (at least if you're running Logstash 1.5 and you might have to crank up the log level) but I also think the Elasticsearch logs will tell you what's wrong.

5

There are clues (error 400, 2 is the wrong type [ paraphrase ]) in the logstash log, but it’s polluted by the dump of the documents. Can I turn off the dump of the document? I am loading a lot of documents in a single file, so it’s hard to find the offending document. I will try to find the document by creating separate files for each document.
I don’t know where the elasticsearch logs are yet. I’m running in esvm dev mode from kibana. I don’t see anything on the grunt console.

John

6

Here's my stackoverflow post:

I have a jsonix JSON schema, here: Example X3D JSON Schema

I'd like to convert the JSON schema to an ElasticSearth type mapping so I can upload my JSON documents into ElasticSearch without type conflicts. I'm not sure what the next step is, and I may go with Mongoose if I can't find a solution (one next step is to hand craft the schema).