[PPL-Lang]PPL support json_delete,json_append functions

[YANG-DB]

Is your feature request related to a problem?
Continue Supporting the PPL Json grammar with adding the next functions:

• json_delete(object,[keys list])

  • Removes one or more keys and their corresponding values from the specified JSON object.
    • objects: the Json object is the field being evaluated
    • [keys list]: keys is the list of keys to remove from the object

• json_append(object, [(path key, value) list])

  • Adds value to an array, json_append tails the value to the arrays into their corespondent path.
    • objects: the Json object is the field being evaluated
    • [(path key, values) list]): path key is the json path name of the array and the value is the object be added

Each designates an array or value within the object json document. The json_extend function adds the values of the corresponding after the last value of the array designated by the .

What solution would you like?

support the following json_delete

Example:

    os> source=people | eval deleted = json_delete({"a":"valueA", "b":"valueB"}, ["a"]) 
    fetched rows / total rows = 1/1
    +----------------------------------+
    | deleted                          |
    +----------------------------------+
    | {"b": "valueB" }                 |
    +----------------------------------+

    os> source=people | eval  eval deleted = json_delete({"a":[{"b":1},{"b":2},{"c":3}]}, ["a.b"])
    fetched rows / total rows = 1/1
    +-----------------------------------------------------------+
    | deleted                                                   |
    +-----------------------------------------------------------+
    | {"a":[{"c":3}] }                                          |
    +-----------------------------------------------------------+

    os> source=people | eval `no_action` =  json_delete({"a":[{"b":1},{"b":2},{"c":3}]}, ["b.b"])
    fetched rows / total rows = 1/1
    +-----------------------------------+
    | no_action                         |
    +-----------------------------------+
    | {"a":[{"b":1},{"b":2},{"c":3}]}   |
    +-----------------------------------+

support the following json_append

    os> source=people | eval append = json_append(`{"a":["valueA", "valueB"]}`, ["a","valueC"]) 
    fetched rows / total rows = 1/1
    +-------------------------------------------------+
    | append                                          |
    +-------------------------------------------------+
    | {"a":["valueA", "valueB", "valueC"]}            |
    +-------------------------------------------------+

    os> source=people | eval append = json_append(`{"a":["valueA", "valueB"]}`, {"a":["valueC"]}) 
    fetched rows / total rows = 1/1
    +-----------------------------------------------+
    | append                                        |
    +-----------------------------------------------+
    | {"a":["valueA", "valueB", ["valueC"]]}        |
    +-----------------------------------------------+

    os> source=people | eval append = json_append(`{"root":{ "a":["valueA", "valueB"]}}`, {"root.a":"valueC"}) 
    fetched rows / total rows = 1/1
    +-----------------------------------------------+
    | append                                        |
    +-----------------------------------------------+
    |{"root": {"a":["valueA", "valueB", "valueC"]}} |
    +-----------------------------------------------+

Do you have any additional context?

• [FEATURE] Enhance JSON array functions #863
• [FEATURE]Add PPL JSON extended functions support #667

[LantaoJin]

 os> source=people | eval deleted = json_delete({"a":"valueA", "b":"valueB"}, ["a"]) 
    fetched rows / total rows = 1/1
    +----------------------------------+
    | deleted                          |
    +----------------------------------+
    | {"a": "valueA" }                 |
    +----------------------------------+

Should be

 os> source=people | eval deleted = json_delete({"a":"valueA", "b":"valueB"}, ["a"]) 
    fetched rows / total rows = 1/1
    +----------------------------------+
    | deleted                          |
    +----------------------------------+
    | {"b": "valueB" }                 |
    +----------------------------------+

os> source=people | eval append = json_append(`{"a":["valueA", "valueB"]}`, {"a":["valueC"]}) 
    fetched rows / total rows = 1/1
    +-----------------------------------------------+
    | append                                        |
    +-----------------------------------------------+
    | {"a":["valueA", "valueB", ["valueC"]]}        |
    +-----------------------------------------------+

We might be hard to support this case since the JSON ARRAY was implemented by Spark ARRAY which all elements in it requires the same type.
So does case

    os> source=people | eval extend = json_extend(`{"a":["valueA", "valueB"]}`, {"b":["valueC","valueD"]}) 
    fetched rows / total rows = 1/1
    +-------------------------------------------------------------+
    | extend                                                      |
    +-------------------------------------------------------------+
    | {"a":["valueA", "valueB", {"b":"valueC"}, {"b":"valueD"}]}  |
    +-------------------------------------------------------------+