UDF EXTRACTJSONFIELD does not work in combination of TRANSFORM

[stefanfrehse]

Describe the bug
Using the UDF EXTRACTJSONFIELD does not correctly parse JSON within a TRANSFORM lambda call while taking the path variable from the TRANSFORM argument.

To Reproduce
Steps to reproduce the behavior, include:

• ksqlDb: 0.20.0 probably earlier as well
• Apply the following statements demonstrates the faulty behavior:

set 'auto.offset.reset'='earliest';
CREATE STREAM STREAM_A  (
    "key" VARCHAR KEY,
    "value" VARCHAR,
    "extractor" ARRAY<VARCHAR>
) WITH(KAFKA_TOPIC = 'test-topic-transform', VALUE_FORMAT = 'JSON', PARTITIONS = 1, REPLICAS = 3);

INSERT INTO STREAM_A ("key", "value", "extractor") VALUES( 'key_1', '{\"sensor_1\": 10, \"sensor_2\": 15}', ARRAY['.sensor_1', '.sensor_2']);
INSERT INTO STREAM_A ("key", "value", "extractor") VALUES( 'key_2', '{\"sensor_1\": 3, \"sensor_2\": 11}', ARRAY['.sensor_1', '.sensor_2']);

SELECT
    "key",
    "value",
    "extractor",
    TRANSFORM("extractor", (data) => EXTRACTJSONFIELD("value", ('$' + data))) AS "unpackedValue"
    FROM STREAM_A
EMIT CHANGES;

which results currently in [10,10], and, [3,3]

Expected behavior
The result of the query should be: [10,15], and, [3,11].

[spena]

Issue fixed here - #8123