[oracle] Splitting comma separated string in a PL/SQL stored proc

I've CSV string 100.01,200.02,300.03 which I need to pass to a PL/SQL stored procedure in Oracle. Inside the proc,I need to insert these values in a Number column in the table.

For this, I got a working approach from over here:

How to best split csv strings in oracle 9i

[2) Using SQL's connect by level.].

Now,I've another requirement. I need to pass 2 CSV strings[equal in length] as input to PL/SQL stored proc.And, I need to break this string and insert each value from two CSV strings into two different columns in the table.Could you please let me know how to go about it?

Example of CSV inputs: mystring varchar2(2000):='0.75, 0.64, 0.56, 0.45';

myAmount varchar2(2000):= '0.25, 0.5, 0.65, 0.8';

myString values would go into Column A and myAmount values into Column B in the table.

Could you please let me know how to achieve this?

Thanks.

This question is related to oracle plsql tokenize

The answer is


This should do what you are looking for.. It assumes your list will always be just numbers. If that is not the case, just change the references to DBMS_SQL.NUMBER_TABLE to a table type that works for all of your data:

CREATE OR REPLACE PROCEDURE insert_from_lists(
    list1_in IN VARCHAR2,
    list2_in IN VARCHAR2,
    delimiter_in IN VARCHAR2 := ','
)
IS 
    v_tbl1 DBMS_SQL.NUMBER_TABLE;
    v_tbl2 DBMS_SQL.NUMBER_TABLE;

    FUNCTION list_to_tbl
    (
        list_in IN VARCHAR2
    )
    RETURN DBMS_SQL.NUMBER_TABLE
    IS
        v_retval DBMS_SQL.NUMBER_TABLE;
    BEGIN

        IF list_in is not null
        THEN
            /*
            || Use lengths loop through the list the correct amount of times,
            || and substr to get only the correct item for that row
            */
            FOR i in 1 .. length(list_in)-length(replace(list_in,delimiter_in,''))+1
            LOOP
                /*
                || Set the row = next item in the list
                */
                v_retval(i) := 
                        substr (
                            delimiter_in||list_in||delimiter_in,
                            instr(delimiter_in||list_in||delimiter_in, delimiter_in, 1, i  ) + 1,
                            instr (delimiter_in||list_in||delimiter_in, delimiter_in, 1, i+1) - instr (delimiter_in||list_in||delimiter_in, delimiter_in, 1, i) -1
                        );
            END LOOP;
        END IF;

        RETURN v_retval;

    END list_to_tbl;
BEGIN 
   -- Put lists into collections
   v_tbl1 := list_to_tbl(list1_in);
   v_tbl2 := list_to_tbl(list2_in);

   IF v_tbl1.COUNT <> v_tbl2.COUNT
   THEN
      raise_application_error(num => -20001, msg => 'Length of lists do not match');
   END IF;

   -- Bulk insert from collections
   FORALL i IN INDICES OF v_tbl1
      insert into tmp (a, b)
      values (v_tbl1(i), v_tbl2(i));

END insert_from_lists; 

As for the connect by use case, this approach should work for you:

select regexp_substr('SMITH,ALLEN,WARD,JONES','[^,]+', 1, level)
from dual
connect by regexp_substr('SMITH,ALLEN,WARD,JONES', '[^,]+', 1, level) is not null;

I am not sure if this fits your oracle version. On my 10g I can use pipelined table functions:

set serveroutput on

create type number_list as table of number;

-- since you want this solution
create or replace function split_csv (i_csv varchar2) return number_list pipelined 
  is 
    mystring varchar2(2000):= i_csv;
  begin
    for r in
    ( select regexp_substr(mystring,'[^,]+',1,level) element
        from dual
     connect by level <= length(regexp_replace(mystring,'[^,]+')) + 1
    )
    loop
      --dbms_output.put_line(r.element);
      pipe row(to_number(r.element, '999999.99'));
    end loop;
  end;
/

insert into foo
select column_a,column_b from 
  (select column_value column_a, rownum rn from table(split_csv('0.75, 0.64, 0.56, 0.45'))) a 
 ,(select column_value column_b, rownum rn from table(split_csv('0.25, 0.5, 0.65, 0.8'))) b
 where a.rn = b.rn
;

create or replace procedure pro_ss(v_str varchar2) as
v_str1 varchar2(100); 
v_comma_pos number := 0;    
v_start_pos number := 1;
begin             
    loop        
    v_comma_pos := instr(v_str,',',v_start_pos);   
    if  v_comma_pos = 0 then     
      v_str1 := substr(v_str,v_start_pos);  
      dbms_output.put_line(v_str1);    
      exit;
      end if;    
    v_str1 := substr(v_str,v_start_pos,(v_comma_pos - v_start_pos)); 
    dbms_output.put_line(v_str1);       
    v_start_pos := v_comma_pos + 1;    
    end loop; 
end;
/

call pro_ss('aa,bb,cc,dd,ee,ff,gg,hh,ii,jj');

outout: aa bb cc dd ee ff gg hh ii jj


Many good solutions have been provided already. However, if he text is provided in a very simple comma delimited format or similar, and speed is of importance, then I have for you a solution with a TABLE function (in PL/SQL). I have also provided a rundown of some other solutions.

Please see more on the Blog Entry on Parsing a CSV into multiple columns.


CREATE OR REPLACE PROCEDURE insert_into (
   p_errcode        OUT   NUMBER,
   p_errmesg        OUT   VARCHAR2,
   p_rowsaffected   OUT   INTEGER
)
AS
   v_param0   VARCHAR2 (30) := '0.25,2.25,33.689, abc, 99';
   v_param1   VARCHAR2 (30) := '2.65,66.32, abc-def, 21.5';
BEGIN
   FOR i IN (SELECT COLUMN_VALUE
               FROM TABLE (SPLIT (v_param0, ',')))
   LOOP
      INSERT INTO tempo
                  (col1
                  )
           VALUES (i.COLUMN_VALUE
                  );
   END LOOP;

   FOR i IN (SELECT COLUMN_VALUE
               FROM TABLE (SPLIT (v_param1, ',')))
   LOOP
      INSERT INTO tempo
                  (col2
                  )
           VALUES (i.COLUMN_VALUE
                  );
   END LOOP;
END;

Here is a good solution:

FUNCTION comma_to_table(iv_raw IN VARCHAR2) RETURN dbms_utility.lname_array IS
   ltab_lname dbms_utility.lname_array;
   ln_len     BINARY_INTEGER;
BEGIN
   dbms_utility.comma_to_table(list   => iv_raw
                              ,tablen => ln_len
                              ,tab    => ltab_lname);
   FOR i IN 1 .. ln_len LOOP
      dbms_output.put_line('element ' || i || ' is ' || ltab_lname(i));
   END LOOP;
   RETURN ltab_lname;
END;

Source: CSV - comma separated values - and PL/SQL (link no longer valid)


I use apex_util.string_to_table to parse strings, but you can use a different parser if you wish. Then you can insert the data as in this example:

declare
  myString varchar2(2000) :='0.75, 0.64, 0.56, 0.45';
  myAmount varchar2(2000) :='0.25, 0.5, 0.65, 0.8';
  v_array1 apex_application_global.vc_arr2;
  v_array2 apex_application_global.vc_arr2;
begin

  v_array1 := apex_util.string_to_table(myString, ', ');
  v_array2 := apex_util.string_to_table(myAmount, ', ');

  forall i in 1..v_array1.count
     insert into mytable (a, b) values (v_array1(i), v_array2(i));
end;

Apex_util is available from Oracle 10G onwards. Prior to this it was called htmldb_util and was not installed by default. If you can't use that you could use the string parser I wrote many years ago and posted here.


Examples related to oracle

concat yesterdays date with a specific time ORA-28001: The password has expired how to modify the size of a column How to create a blank/empty column with SELECT query in oracle? Find the number of employees in each department - SQL Oracle Query to display all tablespaces in a database and datafiles When or Why to use a "SET DEFINE OFF" in Oracle Database How to insert date values into table error: ORA-65096: invalid common user or role name in oracle In Oracle SQL: How do you insert the current date + time into a table?

Examples related to plsql

How can you tell if a value is not numeric in Oracle? Why do I get PLS-00302: component must be declared when it exists? Split function in oracle to comma separated values with automatic sequence PLS-00428: an INTO clause is expected in this SELECT statement What is the max size of VARCHAR2 in PL/SQL and SQL? PLS-00201 - identifier must be declared Default Values to Stored Procedure in Oracle How to call Oracle MD5 hash function? Proper way of checking if row exists in table in PL/SQL block PLS-00103: Encountered the symbol when expecting one of the following:

Examples related to tokenize

How to get rid of punctuation using NLTK tokenizer? How do I tokenize a string sentence in NLTK? Splitting string into multiple rows in Oracle Parse (split) a string in C++ using string delimiter (standard C++) How to use stringstream to separate comma separated strings Split string with PowerShell and do something with each token Splitting comma separated string in a PL/SQL stored proc Convert comma separated string to array in PL/SQL Is there a function to split a string in PL/SQL? How to split a string in shell and get the last field