What happened for NDJSON support on CLI?

[thomas-k-cameron]

Describe the bug
When I tried to run this script, it shown an error.
I believe that it worked before.

CREATE EXTERNAL TABLE IF NOT EXISTS asdf STORED AS NDJSON LOCATION 'some.ndjson';
SELECT * FROM asdf LIMIT 5;

Execution("Unable to find factory for NDJSON")

contents of the file is

{"asdf": 1}
{"asdf": 2}

To Reproduce
Use that file and run that script.

Expected behavior
It should create a table.

Additional context

datafusion-cli 14.0.0

When I tried it with STORED AS JSON it still didn't work.

CREATE EXTERNAL TABLE IF NOT EXISTS asdf STORED AS JSON LOCATION 'some.ndjson';
SELECT * FROM asdf LIMIT 5;

[andygrove]

There have been changes in how table providers are registered so this is likely a regression. It would be good to get some regression tests added when we fix this.

I am guessing that the fix is simply to add NDJSON to this code in datafusion-cli:

fn create_runtime_env() -> Result<RuntimeEnv> {
    let mut table_factories: HashMap<String, Arc<dyn TableProviderFactory>> =
        HashMap::new();
    table_factories.insert(
        "csv".to_string(),
        Arc::new(ListingTableFactory::new(FileType::CSV)),
    );
    table_factories.insert(
        "parquet".to_string(),
        Arc::new(ListingTableFactory::new(FileType::PARQUET)),
    );
    table_factories.insert(
        "avro".to_string(),
        Arc::new(ListingTableFactory::new(FileType::AVRO)),
    );
    table_factories.insert(
        "json".to_string(),
        Arc::new(ListingTableFactory::new(FileType::JSON)),
    );

[thomas-k-cameron]

What do you think about adding several test-cases for the datafusion-cli to check if some features works as intended?

Is there a directory intended for adding test-cases for the cli?(I couldn't find it/)

[timvw]

Seems to have been gone missing in #1010

[timvw]

Seems that this now available as regular "json" format:
https://github.com/apache/arrow-datafusion/blob/master/datafusion/core/src/datasource/file_format/json.rs#L46

The following works on my machine:

DataFusion CLI v14.0.0
❯ create external table x stored as json location 'datafusion/core/tests/jsons/1.json';
0 rows in set. Query took 0.021 seconds.
❯ select * from x;
+-----+----------------+---------------+------+
| a   | b              | c             | d    |
+-----+----------------+---------------+------+
| 1   | [2, 1.3, -6.1] | [false, true] | 4    |
| -10 | [2, 1.3, -6.1] | [true, true]  | 4    |
| 2   | [2, , -6.1]    | [false, ]     | text |
|     |                |               |      |
+-----+----------------+---------------+------+
4 rows in set. Query took 0.030 seconds.

I have updated PR-4427 such that both JSON and NDJSON are registered by default

[thomas-k-cameron]

I'm closing this issue because the problem has been fixed with PR-4427.