Csv upload -- suggestion needed with non-English character in csv file

Question

<p>Hi All,</p>
I have a process which uploads a csv file into a table. It works with the normal english characters. In case of non-English characters in the csv file it doesn't populate the actual columns.
My csv file content is
<p></p>First Name | Middle Name | Last Name
<p><span style="background-color: #FF0000">José</span> | # | Reema</p>
<p>Sam | # | Peter</p>
<p>Out put is coming like : (the last name is coming as blank )</p>
First Name | Middle Name | Last Name
<p><span style="background-color: #FF0000">Jos鬣</span> | Reema | <span style="background-color: #FF0000"> blank </span></p>
<p>Sam | # | Peter</p>
http://apex.oracle.com/pls/otn/f?p=53121:1
workspace- gil_dev
user- apex
password- apex12
Thanks for your help.
Manish

Answer 1

Manish,
PROCEDURE csv_to_array (
      -- Utility to take a CSV string, parse it into a PL/SQL table
      -- Note that it takes care of some elements optionally enclosed
      -- by double-quotes.
      p_csv_string   IN       VARCHAR2,
      p_array        OUT      wwv_flow_global.vc_arr2,
      p_separator    IN       VARCHAR2 := ';'
   IS
      l_start_separator   PLS_INTEGER    := 0;
      l_stop_separator    PLS_INTEGER    := 0;
      l_length            PLS_INTEGER    := 0;
      l_idx               BINARY_INTEGER := 0;
      l_quote_enclosed    BOOLEAN        := FALSE;
      l_offset            PLS_INTEGER    := 1;
   BEGIN
      l_length := NVL (LENGTH (p_csv_string), 0);
      IF (l_length <= 0)
      THEN
         RETURN;
      END IF;
      LOOP
         l_idx := l_idx + 1;
         l_quote_enclosed := FALSE;
         IF SUBSTR (p_csv_string, l_start_separator + 1, 1) = '"'
         THEN
            l_quote_enclosed := TRUE;
            l_offset := 2;
            l_stop_separator :=
                   INSTR (p_csv_string, '"', l_start_separator + l_offset, 1);
         ELSE
            l_offset := 1;
            l_stop_separator :=
               INSTR (p_csv_string,
                      p_separator,
                      l_start_separator + l_offset,
                      1
         END IF;
         IF l_stop_separator = 0
         THEN
            l_stop_separator := l_length + 1;
         END IF;
         p_array (l_idx) :=
            (SUBSTR (p_csv_string,
                     l_start_separator + l_offset,
                     (l_stop_separator - l_start_separator - l_offset
         EXIT WHEN l_stop_separator >= l_length;
         IF l_quote_enclosed
         THEN
            l_stop_separator := l_stop_separator + 1;
         END IF;
         l_start_separator := l_stop_separator;
      END LOOP;
   END csv_to_array;and
PROCEDURE get_records (p_clob IN CLOB, p_records OUT varchar2_t)
   IS
      l_record_separator   VARCHAR2 (2) := CHR (13) || CHR (10);
      l_last               INTEGER;
      l_current            INTEGER;
   BEGIN
      -- SIf HTMLDB has generated the file,
      -- it will be a Unix text file. If user has manually created the file, it
      -- will have DOS newlines.
      -- If the file has a DOS newline (cr+lf), use that
      -- If the file does not have a DOS newline, use a Unix newline (lf)
      IF (NVL (DBMS_LOB.INSTR (p_clob, l_record_separator, 1, 1), 0) = 0)
      THEN
         l_record_separator := CHR (10);
      END IF;
      l_last := 1;
      LOOP
         l_current := DBMS_LOB.INSTR (p_clob, l_record_separator, l_last, 1);
         EXIT WHEN (NVL (l_current, 0) = 0);
         p_records (p_records.COUNT + 1) :=
            REPLACE (DBMS_LOB.SUBSTR (p_clob, l_current - l_last, l_last),
         l_last := l_current + LENGTH (l_record_separator);
      END LOOP;
   END get_records;Denes Kubicek
http://deneskubicek.blogspot.com/
http://www.opal-consulting.de/training
http://htmldb.oracle.com/pls/otn/f?p=31517:1
-------------------------------------------------------------------