JSON Массив на JSON объект с использованием KSTREAM или KSQL

Question

У меня есть данные, поступающие в Kafka в следующем формате.

{"WHS":[{"Character Set":"UTF-8","action":"finished","Update-Date-Time":"2020-04-11 09:00:02:25","Number":0,"Abbr":"","Name":"","Name2":"","Country-Code":"","Addr-1":"","Addr-2":"","Addr-3":"","Addr-4":"","City":"","State":""}]}

Я хочу, чтобы они конвертировались следующим образом.

{"Character Set":"UTF-8","action":"finished","Update-Date-Time":"2020-04-11 09:00:02:25","Number":0,"Abbr":"","Name":"","Name2":"","Country-Code":"","Addr-1":"","Addr-2":"","Addr-3":"","Addr-4":"","City":"","State":""}

Я пытался сгладить, используя k sql но k sql пока не поддерживает массивы.

Я пытался сгладить, используя kstream, используя приведенный ниже код.

builder.stream(inputTopic).flatMapValues(Object -> Arrays.asList()).to(outputTopic);

Но он не производит никакого вывода. Любая помощь с этим будет принята с благодарностью.

Answer 1

K SQL / ksqlDB поддерживает массивы. Вот как сделать то, о чем вы просите:

-- Declare the stream
CREATE STREAM TEST1 
    (WHS ARRAY<STRUCT<"action"           VARCHAR
                    , "Update-Date-Time" VARCHAR
                    , "Number"           VARCHAR
                    , "Abbr"             VARCHAR
                    , "Name"             VARCHAR
                    , "Name2"            VARCHAR
                    , "Country-Code"     VARCHAR
                    , "Addr-1"           VARCHAR
                    , "Addr-2"           VARCHAR
                    , "Addr-4"           VARCHAR
                    , "City"             VARCHAR
                    , "State"            VARCHAR>>) 
    WITH (KAFKA_TOPIC ='test1'
         ,VALUE_FORMAT='JSON');

-- Set querying from beginning of the topic
SET 'auto.offset.reset' = 'earliest';

-- Query the array         
ksql> SELECT WHS FROM TEST1 EMIT CHANGES LIMIT 1;
+------------------------------------------------------------------------------------------------------------------------------------------------------+
|WHS                                                                                                                                                   |
+------------------------------------------------------------------------------------------------------------------------------------------------------+
|[{ACTION=finished, Update-Date-Time=2020-04-11 09:00:02:25, NUMBER=0, ABBR=, NAME=, NAME2=, Country-Code=, Addr-1=, Addr-2=, Addr-4=, City=, STATE=}] |
Limit Reached
Query terminated
ksql>         

-- Flatten the array
ksql> SELECT EXPLODE(WHS) FROM TEST1 EMIT CHANGES LIMIT 1;
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|KSQL_COL_0                                                                                                                                                                                                                                                                               |
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|{ACTION=finished, Update-Date-Time=2020-04-11 09:00:02:25, NUMBER=0, ABBR=, NAME=, NAME2=, Country-Code=, Addr-1=, Addr-2=, Addr-4=, City=, STATE=}                                                                                                                                      |
Limit Reached
Query terminated
ksql>

Вы можете записать это в другой поток (topi c):

ksql> CREATE STREAM TEST1_EXPLODE WITH (KAFKA_TOPIC='NEW_TEST1') AS SELECT EXPLODE(WHS) FROM TEST1 EMIT CHANGES;

 Message
-------------------------------------------
 Created query with ID CSAS_TEST1_EXPLODE_155
-------------------------------------------
ksql> PRINT NEW_TEST1;
…
Value format: JSON or KAFKA_STRING
rowtime: 4/27/20 8:28:46 AM UTC, key: <null>, value: {"KSQL_COL_0":{"ACTION":"finished","Update-Date-Time":"2020-04-11 09:00:02:25","NUMBER":"0","ABBR":"","NAME":"","NAME2":"","Country-Code":"","Addr-1":"","Addr-2":"","Addr-4":"","City":"","STATE":""}}

Если вы хотите сгладить Получив структуру вы тоже можете это сделать:

CREATE STREAM TEST1_FLATTENED AS SELECT  EXPLODE(WHS)->"action"           AS "action"           ,
        EXPLODE(WHS)->"Update-Date-Time" AS "Update-Date-Time" ,
        EXPLODE(WHS)->"Number"           AS "Number"           ,
        EXPLODE(WHS)->"Abbr"             AS "Abbr"             ,
        EXPLODE(WHS)->"Name"             AS "Name"             ,
        EXPLODE(WHS)->"Name2"            AS "Name2"            ,
        EXPLODE(WHS)->"Country-Code"     AS "Country-Code"     ,
        EXPLODE(WHS)->"Addr-1"           AS "Addr-1"           ,
        EXPLODE(WHS)->"Addr-2"           AS "Addr-2"           ,
        EXPLODE(WHS)->"Addr-4"           AS "Addr-4"           ,
        EXPLODE(WHS)->"City"             AS "City"             ,
        EXPLODE(WHS)->"State"            AS "State"
    FROM TEST1 EMIT CHANGES;

ksql> PRINT TEST1_FLATTENED;
…
Value format: JSON or KAFKA_STRING
rowtime: 4/27/20 8:28:46 AM UTC, key: <null>, value: {"action":"finished","Update-Date-Time":"2020-04-11 09:00:02:25","Number":"0","Abbr":"","Name":"","Name2":"","Country-Code":"","Addr-1":"","Addr-2":"","Addr-4":"","City":"","State":""}

Answer 2

Arrays.asList () создает только пустой список.

Добавьте фактические логи c, чтобы получить массив из ввода и преобразовать его в коллекцию (например, ArrayList), которая реализует Iterable, здесь я пытаюсь использовать flatMapValues ​​с Джексоном:

        builder.stream(inputTopic).flatMapValues((ValueMapper<JsonNode, Iterable<JsonNode>>) value -> {
            ArrayNode arrayNode = (ArrayNode) value.get("WHS");
            return arrayNode::iterator;
        });