Nested column branch had multiple children

[asfimport]

from pyarrow import json
import pyarrow.parquet as pq

r = json.read_json('example.jl')
pq.write_table(r, 'example.parquet')

Doing the above operation resulting in ArrowInvalid: Nested column branch had multiple children

Posting it here as per the request from #4045 (comment)

The sample schema looks like this

package_version: string
source_version: string
uuid: string
_type: string
position: struct<ais_type: string, course: double, draught: double, draught_raw: null, heading: double, lat: double, lon: double, nav_state: int64, received_time: timestamp[s], speed: double>
 child 0, ais_type: string
 child 1, course: double
 child 2, draught: double
 child 3, draught_raw: null
 child 4, heading: double
 child 5, lat: double
 child 6, lon: double
 child 7, nav_state: int64
 child 8, received_time: timestamp[s]
 child 9, speed: double
provider_name: string
vessel: struct<beam: null, build_year: null, call_sign: string, dead_weight: null, dwt: null, flag_code: null, flag_name: string, gross_tonnage: null, imo: string, length: null, mmsi: string, name: string, type: null, vessel_type: string>
 child 0, beam: null
 child 1, build_year: null
 child 2, call_sign: string
 child 3, dead_weight: null
 child 4, dwt: null
 child 5, flag_code: null
 child 6, flag_name: string
 child 7, gross_tonnage: null
 child 8, imo: string
 child 9, length: null
 child 10, mmsi: string
 child 11, name: string
 child 12, type: null
 child 13, vessel_type: string
source_provider: string

 

 

Reporter: harikrishnan

Related issues:

• [C++][Parquet] Read and write nested Parquet data with a mix of struct and list nesting levels (is duplicated by)

Original Issue Attachments:

• SampleRecord.jl

_{Note: This issue was originally created as ARROW-6737. Please see the migration documentation for further details.}

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
[~harish1792] would you be able to provide a reproducible example? (so we can run the code and investigate the issue)
For example, a small json file that shows the problem?

[asfimport]

Wes McKinney / @wesm:
Is this ARROW-1644?

[asfimport]

harikrishnan:
@jorisvandenbossche Please use the attached file for testing. The attached file is not the same one as the schema I posted above. But this would result in the same error as I posted above. Let me know if you need any more info!

[asfimport]

harikrishnan:
Not sure https://jira.apache.org/jira/browse/ARROW-6760# this is also related to this. But facing this when I am trying to do a similar option with a slighty different Json format

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
Thanks for providing the sample file. This is indeed a duplicate of ARROW-1644. Nested lists/structs are currently not yet supported in the Arrow parquet IO implementation.

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
I noticed that reading this file on master actually gives problems, while it works on 0.14.1, so opened ARROW-6762 for that.