skip to Main Content

The issue when I don’t use AGG type I get a ‘ora-01422: exact fetch returns more than requested number of rows using a json_object put into clob so using JSON_ARRAYAGG works I just haven’t been able to figure out how to format the query properly to return the JSON object that I need.

I’m not a Oracle guru, or database admin, so that’s probably the issue, any help from you guys would be great. In the meantime I will continue to play with it and maybe I can find my own solution.

This is my Stored Procedure Package Body:

  PROCEDURE getCities
    (
    v_province_fips IN geo_counties.FIPS_CODE%TYPE,
    v_city_like in geo_cities.NAME%TYPE, 
    p_out OUT CLOB
    )
AS

    v_country_code geo_cities.country_code%TYPE;
    v_fips_code geo_counties.FIPS_CODE%TYPE;

BEGIN

    v_country_code := SUBSTR('' || v_province_fips || '',1,2);
    
            
        SELECT JSON_ARRAYAGG(
        JSON_OBJECTAGG('CITY_SEARCH' VALUE
            JSON_OBJECT(
                'GEONAME_ID_CITY' VALUE c.GEONAME_ID,
                'NAME' VALUE c.NAME,
                'ASCII_NAME' VALUE c.ASCII_NAME,
                'LATITUDE' VALUE c.LATITUDE,
                'LONGITUDE' VALUE c.LONGITUDE,
                'STATE_PROV_NAME' VALUE a.NAME,
                'GEONAME_ID_COUNTY' VALUE b.GEONAME_ID,
                'COUNTY_NAME' VALUE b.NAME,
                'COUNTY_ASCII_NAME' VALUE b.ASCII_NAME,
                'STATE_PROV' VALUE p.NAME,
                'COUNTRY_CODE' VALUE c.COUNTRY_CODE,
                'COUNTY_LATITUDE' VALUE b.LATITUDE,
                'COUNTY_LONGITUDE' VALUE b.LONGITUDE,
                'FIPS_CODE' VALUE b.FIPS_CODE 
                )
                )
                RETURNING CLOB)

                
                INTO p_out
                FROM geo_cities  c
                JOIN  geo_counties b
                ON (b.COUNTRY_CODE = c.COUNTRY_CODE)
                JOIN GEO_ADMIN1_CODES_ASCII a
                ON (a.FIPS_CODE = v_country_code || '.' || SUBSTR('' || b.FIPS_CODE || '',4,2))
                JOIN GEO_PROVINCES p
                ON (p.FIPS_CODE = v_country_code || '.' || SUBSTR('' || b.FIPS_CODE || '',4,2))
                where  LOWER( c.NAME ) like LOWER( '%' || v_city_like  || '%' )
                AND c.COUNTRY_CODE = v_country_code
                AND c.ADMIN_1 =  SUBSTR('' || b.FIPS_CODE || '',4,2)
                AND c.ADMIN_2 = SUBSTR('' || b.FIPS_CODE || '',7)
                AND c.feature_code in ('ADM1','ADM2','PPL','PPLA', 'PPLA2')
                GROUP BY 
                c.GEONAME_ID,
                c.NAME,
                c.ASCII_NAME,
                c.LATITUDE,
                c.LONGITUDE,
                c.COUNTRY_CODE,
                c.ADMIN_1,
                b.ADMIN_2,
                b.GEONAME_ID,
                b.NAME,
                a.NAME,
                b.ASCII_NAME,
                c.COUNTRY_CODE,
                b.LATITUDE,
                b.LONGITUDE,
                b.POPULATION,
                b.TIME_ZONE,
                b.FIPS_CODE
                ORDER BY c."NAME", b.NAME,c.ADMIN_1;
END getCities;

This returns the following json object:

[
    {
        "CITY_SEARCH": {
            "GEONAME_ID_CITY": 4164138,
            "NAME": "Miami",
            "ASCII_NAME": "Miami",
            "LATITUDE": 25.77427,
            "LONGITUDE": -80.19366,
            "STATE_PROV_NAME": "Florida",
            "GEONAME_ID_COUNTY": 4164238,
            "COUNTY_NAME": "Miami-Dade County",
            "COUNTY_ASCII_NAME": "Miami-Dade County",
            "STATE_PROV": "Florida",
            "COUNTRY_CODE": "US",
            "COUNTY_LATITUDE": 25.60897,
            "COUNTY_LONGITUDE": -80.49867,
            "FIPS_CODE": "US.FL.086"
        }
    },
    {
        "CITY_SEARCH": {
            "GEONAME_ID_CITY": 4164143,
            "NAME": "Miami Beach",
            "ASCII_NAME": "Miami Beach",
            "LATITUDE": 25.79065,
            "LONGITUDE": -80.13005,
            "STATE_PROV_NAME": "Florida",
            "GEONAME_ID_COUNTY": 4164238,
            "COUNTY_NAME": "Miami-Dade County",
            "COUNTY_ASCII_NAME": "Miami-Dade County",
            "STATE_PROV": "Florida",
            "COUNTRY_CODE": "US",
            "COUNTY_LATITUDE": 25.60897,
            "COUNTY_LONGITUDE": -80.49867,
            "FIPS_CODE": "US.FL.086"
        }
    }
]

The query works as expected, it’s the JSON format that is the issue. I want to return a JSON object as follows:

{
    "CITY_SEARCH": [
        {
            "GEONAME_ID_CITY": 4164143,
            "NAME": "Miami Beach",
            "ASCII_NAME": "Miami Beach",
            "LATITUDE": 25.79065,
            "LONGITUDE": -80.13005,
            "STATE_PROV_NAME": "Florida",
            "GEONAME_ID_COUNTY": 4164238,
            "COUNTY_NAME": "Miami-Dade County",
            "COUNTY_ASCII_NAME": "Miami-Dade County",
            "STATE_PROV": "Florida",
            "COUNTRY_CODE": "US",
            "COUNTY_LATITUDE": 25.60897,
            "COUNTY_LONGITUDE": -80.49867,
            "FIPS_CODE": "US.FL.086"
        },
        {
            "GEONAME_ID_CITY": 4164138,
            "NAME": "Miami",
            "ASCII_NAME": "Miami",
            "LATITUDE": 25.77427,
            "LONGITUDE": -80.19366,
            "STATE_PROV_NAME": "Florida",
            "GEONAME_ID_COUNTY": 4164238,
            "COUNTY_NAME": "Miami-Dade County",
            "COUNTY_ASCII_NAME": "Miami-Dade County",
            "STATE_PROV": "Florida",
            "COUNTRY_CODE": "US",
            "COUNTY_LATITUDE": 25.60897,
            "COUNTY_LONGITUDE": -80.49867,
            "FIPS_CODE": "US.FL.086"
        }
    ]
}

2

Answers


  1. You don’t need the objectagg, or the grouping.

    With a simplified demo your query is doing this:

    SELECT JSON_ARRAYAGG(
            JSON_OBJECTAGG('CITY_SEARCH' VALUE
                JSON_OBJECT(
                    'GEONAME_ID_CITY' VALUE c.GEONAME_ID,
                    'NAME' VALUE c.NAME
                )
            )
        )
    FROM c
    GROUP BY c.GEONAME_ID, c.NAME
    
    JSON_ARRAYAGG(JSON_OBJECTAGG(‘CITY_SEARCH’VALUEJSON_OBJECT(‘GEONAME_ID_CITY’VALUEC.GEONAME_ID,’NAME’VALUEC.NAME)))
    [{"CITY_SEARCH":{"GEONAME_ID_CITY":4164138,"NAME":"Miami"}},{"CITY_SEARCH":{"GEONAME_ID_CITY":4164143,"NAME":"Miami Beach"}}]

    or

    [
      {
        "CITY_SEARCH": {
          "GEONAME_ID_CITY": 4164138,
          "NAME": "Miami"
        }
      },
      {
        "CITY_SEARCH": {
          "GEONAME_ID_CITY": 4164143,
          "NAME": "Miami Beach"
        }
      }
    ]
    

    … which is a cut-down version of what you are seeing.

    If you instead have a single object with a value that is an json_arrayagg of objects, then you don’t need (or want) the group-by, and you get:

    SELECT JSON_OBJECT('CITY_SEARCH' VALUE
            JSON_ARRAYAGG(
                JSON_OBJECT(
                    'GEONAME_ID_CITY' VALUE c.GEONAME_ID,
                    'NAME' VALUE c.NAME
                )
            )
        )
    FROM c
    
    JSON_OBJECT(‘CITY_SEARCH’VALUEJSON_ARRAYAGG(JSON_OBJECT(‘GEONAME_ID_CITY’VALUEC.GEONAME_ID,’NAME’VALUEC.NAME)))
    {"CITY_SEARCH":[{"GEONAME_ID_CITY":4164138,"NAME":"Miami"},{"GEONAME_ID_CITY":4164143,"NAME":"Miami Beach"}]}

    or

    {
      "CITY_SEARCH": [
        {
          "GEONAME_ID_CITY": 4164138,
          "NAME": "Miami"
        },
        {
          "GEONAME_ID_CITY": 4164143,
          "NAME": "Miami Beach"
        }
      ]
    }
    

    … which is a cut-down version of what you want to see.

    fiddle

    Login or Signup to reply.
  2. You want to get rid of the JSON_OBJECTAGG and the GROUP BY clause:

    PROCEDURE getCities (
      v_province_fips IN  geo_counties.FIPS_CODE%TYPE,
      v_city_like     IN  geo_cities.NAME%TYPE, 
      p_out           OUT CLOB
    )
    AS
      v_country_code geo_cities.country_code%TYPE := SUBSTR(v_province_fips,1,2);
    BEGIN
      SELECT JSON_OBJECT(
               KEY 'CITY_SEARCH' VALUE JSON_ARRAYAGG(
                 JSON_OBJECT(
                   KEY 'GEONAME_ID_CITY'   VALUE c.GEONAME_ID,
                   KEY 'NAME'              VALUE c.NAME,
                   KEY 'ASCII_NAME'        VALUE c.ASCII_NAME,
                   KEY 'LATITUDE'          VALUE c.LATITUDE,
                   KEY 'LONGITUDE'         VALUE c.LONGITUDE,
                   KEY 'STATE_PROV_NAME'   VALUE a.NAME,
                   KEY 'GEONAME_ID_COUNTY' VALUE b.GEONAME_ID,
                   KEY 'COUNTY_NAME'       VALUE b.NAME,
                   KEY 'COUNTY_ASCII_NAME' VALUE b.ASCII_NAME,
                   KEY 'STATE_PROV'        VALUE p.NAME,
                   KEY 'COUNTRY_CODE'      VALUE c.COUNTRY_CODE,
                   KEY 'COUNTY_LATITUDE'   VALUE b.LATITUDE,
                   KEY 'COUNTY_LONGITUDE'  VALUE b.LONGITUDE,
                   KEY 'FIPS_CODE'         VALUE b.FIPS_CODE 
                 )
                 RETURNING CLOB
               )
               RETURNING CLOB
             )
      INTO   p_out
      FROM   geo_cities  c
             JOIN  geo_counties b
             ON (b.COUNTRY_CODE = c.COUNTRY_CODE)
             JOIN GEO_ADMIN1_CODES_ASCII a
             ON (a.FIPS_CODE = v_country_code || '.' || SUBSTR(b.FIPS_CODE,4,2))
                    JOIN GEO_PROVINCES p
                    ON (p.FIPS_CODE = v_country_code || '.' || SUBSTR(b.FIPS_CODE,4,2))
      where  LOWER( c.NAME ) like LOWER( '%' || v_city_like  || '%' )
      AND    c.COUNTRY_CODE = v_country_code
      AND    c.ADMIN_1 =  SUBSTR(b.FIPS_CODE,4,2)
      AND    c.ADMIN_2 = SUBSTR(b.FIPS_CODE,7)
      AND    c.feature_code in ('ADM1','ADM2','PPL','PPLA', 'PPLA2');
    END getCities;
    

    Also, '' || 'something' is pointless as '' is identical to NULL and NULL || 'something' is simply 'something'.

    Login or Signup to reply.
Please signup or login to give your own answer.
Back To Top
Search