如何将带有逗号分隔符的多个值从csv文件插入表中

Question

我创建了一个流程代码，用于将数据从 CSV 文件导入 Apex 表。

The following code i have used in process:

DECLARE
v_blob_data       BLOB;
v_blob_len        NUMBER;
v_position        NUMBER;
v_raw_chunk       RAW(10000);
v_char      CHAR(1);
c_chunk_len   number       := 1;
v_line        VARCHAR2 (32767)        := NULL;
v_data_array      wwv_flow_global.vc_arr2;
v_line_count number := 0;
BEGIN
-- Read data from wwv_flow_files
select blob_content into v_blob_data
 from wwv_flow_files 
 where CREATED_ON = (select max(CREATED_ON) from wwv_flow_files where lower(CREATED_BY) =lower(:APP_USER));

v_blob_len := dbms_lob.getlength(v_blob_data);
v_position := 1;

-- Read and convert binary to char
WHILE ( v_position <= v_blob_len ) LOOP
 v_raw_chunk := dbms_lob.substr(v_blob_data,c_chunk_len,v_position);
 v_char :=  chr(hex_to_decimal(rawtohex(v_raw_chunk)));
 v_line := v_line || v_char;
 v_position := v_position + c_chunk_len;
-- When a whole line is retrieved
 IF v_char = CHR(10) THEN
 v_line_count := v_line_count + 1;
-- Convert comma to : to use wwv_flow_utilities
   v_line := REPLACE (v_line, ',', ':');
-- Convert each column separated by : into array of data
   v_data_array := wwv_flow_utilities.string_to_table (v_line);
-- Insert data into target table
   IF v_line_count > 1 THEN
   EXECUTE IMMEDIATE 'insert into NON_DYNAMIC_USER_GROUPS_TEMP (ID,WORKSPACES,GROUP_NAME,MEMBERS) values (NON_DYNAMIC_GROUPS_TEMP_SEQ.NEXTVAL,(:1),(:2),(:3))'
   USING  
   v_data_array(1), v_data_array(2), v_data_array(3);    
   end if;
-- Clear out
   v_line := NULL;
  END IF;
 END LOOP;
END;

Here for 3rd column i have to insert multiple values with comma separator from csv file, for ex in v_data_array(3) i have to entered ABINNAYA,BARATH,CHELLA this has to insert in table as ABINNAYA,BARATH,CHELLA.

插入表工作正常，但如果第三列有多个值意味着只有多个值不会被插入。

Thanks

Answer 1

如果您使用 OraOpenSource Utils 包集，则有一种更优雅的方式可以将 CSV 数据插入到表中。

使用它，您可以使用类似于以下的块代码将数据插入表中：

declare 
    v_blob_data blob;
    v_clob_data clob;
  -- variables needed for blob-clob convertion
    v_dest_offset integer := 1;
    v_src_offset integer := 1;
    v_lang_context INTEGER := DBMS_LOB.default_lang_ctx;
    v_warning integer := 0;
    -- one line of the CSV data
    v_line oos_util.tab_vc2_arr;
begin
    -- load CSV data into blob var
    select blob_content into v_blob_data
    from wwv_flow_files 
    where CREATED_ON = (select max(CREATED_ON) from wwv_flow_files where lower(CREATED_BY) =lower(:APP_USER));

    -- converting blob to clob
    dbms_lob.createtemporary(v_clob_data, true);
    dbms_lob.converttoclob(v_clob_data, v_blob_data, DBMS_LOB.LOBMAXSIZE, v_dest_offset, v_src_offset, nls_charset_id('UTF8'), v_lang_context, v_warning);

    -- splitting the CSV to strings and then inserting each row attribute by attribute
    for x in (
        select column_value val from table(oos_util_string.listunagg(p_str => v_clob_data, p_delim => chr(10)))
    ) loop
      v_line := oos_util_string.string_to_table(x.val);
        -- you can apply some datatype conversion here if needed (for example with to_date() in case there are date columns)
      insert into ttt values (v_line(1), v_line(2), v_line(3));
    end loop;
    commit;
end;

首先它将您的数据加载到 BLOB 变量中，然后转换为 CLOB 变量（注意代码页），然后使用 OOS Utils 首先将数据拆分为行，最后按属性将数据插入目标表中。

在进行数据类型转换时要准确。

附：在您原来的帖子中，您真的需要使用动态 SQL 来完成您的任务吗？好像不是这样的。

我不确定我是否回答了你的问题，因为我并不完全清楚你的问题是什么。