即使声明了对象，我也收到组件“COUNT”、“GET_OBJECT”、“GET_ARRAY”必须声明错误

Question

下面是我的存储过程，我尝试使用 pl/sql 语句访问端点并读取 json 响应。我需要从使用 JSON_OBJECT 和 JSON_ARRAY 的响应中提取数据。我使用的是 Oracle 19 及以上版本，并且能够访问 JSON 库。我不确定为什么会遇到 PLS-00302 错误，因为必须声明组件“COUNT”、“GET_OBJECT”、“GET_ARRAY”。

create or replace PROCEDURE ProcessTableauReportMetadata(
 tenant_identifier VARCHAR2
) IS
 -- Declare variables
 offset NUMBER := 0;
 limit NUMBER := 50;
 json_response CLOB;
 totalResults NUMBER;

    -- JSON parsing variables
   json_obj JSON_OBJECT_T;
   items JSON_ARRAY_T;
   item JSON_OBJECT_T;

  -- Data processing variables
   workbook_uuid VARCHAR2(36);
    datasource_uuid VARCHAR2(36);

  BEGIN
       LOOP
      -- Construct endpoint URL with parameters
    endpoint_url := 'https://' || tenant_identifier || '.api.us.devhealtheintent.com/business-    intelligence/v1/reports?biType=TABLEAU&lite=false&offset=' || offset || '&limit=' || limit;

-- Begin HTTP request
req := UTL_HTTP.BEGIN_REQUEST(endpoint_url, 'GET');
UTL_HTTP.SET_TRANSFER_TIMEOUT(100);
UTL_HTTP.SET_HEADER(req, 'Authorization', 'Bearer ' || 'TOKEN'); -- Replace with actual token

-- Execute HTTP request
resp := UTL_HTTP.GET_RESPONSE(req);

-- Read the entire JSON response as text
UTL_HTTP.READ_TEXT(resp, json_response);

-- Parse JSON response using built-in JSON parsing functions
json_obj := JSON_OBJECT_T.PARSE(json_response);
items := json_obj.GET_ARRAY('items');

-- Extract totalResults from JSON response
totalResults := json_obj.GET_NUMBER('totalResults');

-- Process each item in the JSON response
FOR i IN 1 .. items.COUNT LOOP
  item := items.GET_OBJECT(i);

  -- Extract workbook_uuid and associatedSources from the item
  workbook_uuid := item.GET_STRING('id');
  associatedSources := item.GET_ARRAY('associatedSources');

  -- Check for non-empty associatedSources
  IF associatedSources.COUNT > 0 THEN
    datasource_uuid := associatedSources.GET_OBJECT(1).GET_STRING('id');

    -- Update datasource_uuid in TABLEAU_USAGE_WORKBOOK_METADATA table
    UPDATE TABLEAU_USAGE_WORKBOOK_METADATA
    SET DATASOURCE_UUID = datasource_uuid
    WHERE WORKBOOK_UUID = workbook_uuid;
  END IF;
END LOOP;

-- Close the HTTP response
UTL_HTTP.END_RESPONSE(resp);

-- Increment offset for the next API call
offset := offset + limit;

-- Exit the loop if totalResults is reached
EXIT WHEN offset >= totalResults;
 END LOOP;

EXCEPTION
 WHEN OTHERS THEN
-- Handle exceptions
DBMS_OUTPUT.PUT_LINE('Error: ' || SQLERRM);
UTL_HTTP.END_RESPONSE(resp);
END ProcessTableauReportMetadata;

我尝试更改声明 JSON_OBJECT 和 JSON_ARRAY 的方式，但没有成功。

Answer 1

忽略缺少的变量声明，问题在于您将

items

视为 PL/SQL 集合类型，但事实并非如此。您还尝试将作为

json_element_t

返回的结果处理为数组和对象，而不是明确地

treat

处理它们。

您可以在文档中阅读有关 JSON 数据结构以及可用于它们的有效方法的更多信息。

所以代替：

FOR i IN 1 .. items.COUNT LOOP
  item := items.GET_OBJECT(i);

使用（注意数组是从零开始索引的，而不是像 PL/SQL 数组那样从 1 开始索引）：

FOR i IN 0 .. items.GET_SIZE - 1 LOOP
  item := TREAT(items.GET(i) AS JSON_OBJECT_T);

然后代替：

  IF associatedSources.COUNT > 0 THEN
    datasource_uuid := associatedSources.GET_OBJECT(1).GET_STRING('id');

用途：

  IF associatedSources.GET_SIZE > 0 THEN
    datasource_uuid := TREAT(associatedSources.GET(0) AS JSON_OBJECT_T).GET_STRING('id');

跳过http处理，数据提取部分可以这样完成：

-- Process each item in the JSON response
--FOR i IN 1 .. items.COUNT LOOP
FOR i IN 0 .. items.GET_SIZE - 1 LOOP
--  item := items.GET_OBJECT(i);
  item := TREAT(items.GET(i) AS JSON_OBJECT_T);

  -- Extract workbook_uuid and associatedSources from the item
  workbook_uuid := item.GET_STRING('id');
  associatedSources := item.GET_ARRAY('associatedSources');

  -- Check for non-empty associatedSources
--  IF associatedSources.COUNT > 0 THEN
  IF associatedSources.GET_SIZE > 0 THEN
--    datasource_uuid := associatedSources.GET_OBJECT(1).GET_STRING('id');
    datasource_uuid := TREAT(associatedSources.GET(0) AS JSON_OBJECT_T).GET_STRING('id');

    -- Update datasource_uuid in TABLEAU_USAGE_WORKBOOK_METADATA table
    UPDATE TABLEAU_USAGE_WORKBOOK_METADATA
    SET DATASOURCE_UUID = datasource_uuid
    WHERE WORKBOOK_UUID = workbook_uuid;
  END IF;
END LOOP;

小提琴