GeoMesa cannot infer a schema for the output of Avro tools tojson

GeoMesa cannot infer a schema for the output of Avro tools tojson

The attached file was created with:

java -jar avro-tools-1.10.1.jar tojson example-v5.avro > example-v5-tojson.json

Attempting to convert it hits an error:

./bin/geomesa-fs convert /opt/avro/example-v5-tojson.json
INFO No converter defined - will attempt to detect schema from input files
ERROR Could not determine converter from inputs - please specify a converter

Original JIRA Issue: https://geomesa.atlassian.net/browse/GEOMESA-3036

Key: GEOMESA-3036
Type: Story
Priority: Major
Status: In Review
Resolution: Unresolved
Assignee: Adeet Patel
Reporter: James Hughes
Created: Mon, 29 Mar 2021 15:58:57 -0400
Updated: Tue, 6 Feb 2024 14:10:29 -0500

Attachments:

• example-v5-tojson.json (0.7 KB) - uploaded by jnh5y on Mon, 29 Mar 2021 15:57:39 -0400

Comments:

557058:da7f6dfa-f0a1-468b-aec5-2cb61273feb5 (Mon, 29 Mar 2021 16:47:24 -0400):

if you have the avro file, why not just that?

557058:48912df4-0004-4f90-9915-db26684ec006 (Mon, 29 Mar 2021 16:58:25 -0400):

Ah! Good question. In the real world situation, the reporter had an Avro file wanted to ingest it. They tried tojson and its output wasn’t working.

This is to report that the style of output from that tool is a no-good for the GeoMesa SFT inferencer.

Reproduce the failure with the attached example-v5-tojson.json using the geomesa-fs convert command, and compare it with the original Avro input produced by avro-tools tojson. Trace the schema inference path used when no converter is defined. Done means the JSON output is recognized and conversion proceeds with an inferred schema.