ORC-236: Support UNION type in Java Convert tool

[rizaon]

What changes were proposed in this pull request?

This patch add support to convert json to UNION type in orc file. For
example, for schema struct<foo:uniontype<int,string>>, the following
json lines can be parsed into UNION type.

{"foo": {"tag": 0, "value": 1}}
{"foo": {"tag": 1, "value": "testing"}}
{"foo": {"tag": 0, "value": 3}}

Why are the changes needed?

This add a missing support for UNION type in java convert tool.

How was this patch tested?

Manually test against handcrafted json file.

[dongjoon-hyun]

Thank you, @rizaon .

[dongjoon-hyun]

cc @guiyanakuang and @williamhyun , too

[dongjoon-hyun]

Is this applicable to branch-1.7, @rizaon ?

[rizaon]

@dongjoon-hyun If backport is clean, it should be applicable.

[dongjoon-hyun]

Thank you, @rizaon . I set the milestone, v1.7.3.

[guiyanakuang]

It might be nice to make Convert more generic, some existing json data may not have tag and value fields, adding an option for the user to customize the method of getting tag and value would be great. : )

[rizaon]

Hello @guiyanakuang, thanks for your feedback!
Does it help if I add parameters to specify the tag and value json-key? Maybe, --union_tag and --union_key like this?

java -jar ./java/tools/target/orc-tools-1.8.0-SNAPSHOT-uber.jar convert -o sample.orc \
  -s "struct<foo:uniontype<int,string>>" --union_tag="tag" --union_value="value" sample.json

[guiyanakuang]

@rizaon. I think it's good that this can handle simple situations. If there is complex json data we can keep upgrading

[dongjoon-hyun]

I switched the milestone to 1.8.0 since 1.7.3 will be released in two weeks and this is a new improvement which is not a blocker for https://github.com/apache/orc/milestone/4?closed=1 .

[guiyanakuang]

+1 LGTM

[dongjoon-hyun]

+1, LGTM. Thank you, @rizaon and @guiyanakuang .
Also, cc @williamhyun because he is the release manager of Apache ORC 1.7.4.

Merged to main/1.7.

[dongjoon-hyun]

@rizaon . I added you to the Apache ORC contributor group and assigned ORC-236 to you.
Welcome to the Apache ORC community!

[rizaon]

Thank you, @dongjoon-hyun @guiyanakuang !