LISTAGG in Oracle to return distinct values

Question

I am trying to use the LISTAGG function in Oracle. I would like to get only the distinct values for that column. Is there a way in which I can get only the distinct values without creating a function or a procedure?

  col1  col2 Created_by
   1     2     Smith 
   1     2     John 
   1     3     Ajay 
   1     4     Ram 
   1     5     Jack 

I need to select col1 and the LISTAGG of col2 (column 3 is not considered). When I do that, I get something like this as the result of LISTAGG: [2,2,3,4,5]

I need to remove the duplicate '2' here; I need only the distinct values of col2 against col1.

Answer 1

The simplest way to handle multiple listagg's is to use 1 WITH (subquery factor) per column containing a listagg of that column from a select distinct:

    WITH tab AS 
    (           
        SELECT 1 as col1, 2 as col2, 3 as col3, 'Smith' as created_by FROM dual
        UNION ALL SELECT 1 as col1, 2 as col2, 3 as col3,'John'  as created_by FROM dual
        UNION ALL SELECT 1 as col1, 3 as col2, 4 as col3,'Ajay'  as created_by FROM dual
        UNION ALL SELECT 1 as col1, 4 as col2, 4 as col3,'Ram'   as created_by FROM dual
        UNION ALL SELECT 1 as col1, 5 as col2, 6 as col3,'Jack'  as created_by FROM dual
    )
    , getCol2 AS
    (
        SELECT  DISTINCT col1, listagg(col2,',') within group (order by col2)  over (partition by col1) AS col2List
        FROM ( SELECT DISTINCT col1,col2 FROM tab)
    )
    , getCol3 AS
    (
        SELECT  DISTINCT col1, listagg(col3,',') within group (order by col3)  over (partition by col1) AS col3List
        FROM ( SELECT DISTINCT col1,col3 FROM tab)
    )
    select col1,col2List,col3List
    FROM getCol2
    JOIN getCol3
    using (col1)

Which gives:

col1  col2List  col3List
1     2,3,4,5   3,4,6

Answer 2

If you do not need a particular order of concatenated values, and the separator can be a comma, you can do:

select col1, stragg(distinct col2)
  from table
 group by col1

Answer 3

I neded a DISTINCT version of this and got this one working out.

RTRIM(REGEXP_REPLACE(
                       (value, ', ') WITHIN GROUP( ORDER BY value)), 
                            '([^ ]+)(, \1)+','\1'),', ') 

Answer 4

If the intent is to apply this transformation to multiple columns, I have extended a_horse_with_no_name's solution:

SELECT * FROM
(SELECT LISTAGG(GRADE_LEVEL, ',') within group(order by GRADE_LEVEL) "Grade Levels" FROM (select distinct GRADE_LEVEL FROM Students) t)                     t1,
(SELECT LISTAGG(ENROLL_STATUS, ',') within group(order by ENROLL_STATUS) "Enrollment Status" FROM (select distinct ENROLL_STATUS FROM Students) t)          t2,
(SELECT LISTAGG(GENDER, ',') within group(order by GENDER) "Legal Gender Code" FROM (select distinct GENDER FROM Students) t)                               t3,
(SELECT LISTAGG(CITY, ',') within group(order by CITY) "City" FROM (select distinct CITY FROM Students) t)                                                  t4,
(SELECT LISTAGG(ENTRYCODE, ',') within group(order by ENTRYCODE) "Entry Code" FROM (select distinct ENTRYCODE FROM Students) t)                             t5,
(SELECT LISTAGG(EXITCODE, ',') within group(order by EXITCODE) "Exit Code" FROM (select distinct EXITCODE FROM Students) t)                                 t6,
(SELECT LISTAGG(LUNCHSTATUS, ',') within group(order by LUNCHSTATUS) "Lunch Status" FROM (select distinct LUNCHSTATUS FROM Students) t)                     t7,
(SELECT LISTAGG(ETHNICITY, ',') within group(order by ETHNICITY) "Race Code" FROM (select distinct ETHNICITY FROM Students) t)                              t8,
(SELECT LISTAGG(CLASSOF, ',') within group(order by CLASSOF) "Expected Graduation Year" FROM (select distinct CLASSOF FROM Students) t)                     t9,
(SELECT LISTAGG(TRACK, ',') within group(order by TRACK) "Track Code" FROM (select distinct TRACK FROM Students) t)                                         t10,
(SELECT LISTAGG(GRADREQSETID, ',') within group(order by GRADREQSETID) "Graduation ID" FROM (select distinct GRADREQSETID FROM Students) t)                 t11,
(SELECT LISTAGG(ENROLLMENT_SCHOOLID, ',') within group(order by ENROLLMENT_SCHOOLID) "School Key" FROM (select distinct ENROLLMENT_SCHOOLID FROM Students) t)       t12,
(SELECT LISTAGG(FEDETHNICITY, ',') within group(order by FEDETHNICITY) "Federal Race Code" FROM (select distinct FEDETHNICITY FROM Students) t)                         t13,
(SELECT LISTAGG(SUMMERSCHOOLID, ',') within group(order by SUMMERSCHOOLID) "Summer School Key" FROM (select distinct SUMMERSCHOOLID FROM Students) t)                               t14,
(SELECT LISTAGG(FEDRACEDECLINE, ',') within group(order by FEDRACEDECLINE) "Student Decl to Prov Race Code" FROM (select distinct FEDRACEDECLINE FROM Students) t)          t15

This is Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production.
I was unable to use STRAGG because there is no way to DISTINCT and ORDER.

Performance scales linearly, which is good, since I am adding all columns of interest. The above took 3 seconds for 77K rows. For just one rollup, .172 seconds. I do with there was a way to distinctify multiple columns in a table in one pass.

Answer 5

I implemented this stored function :

CREATE TYPE LISTAGG_DISTINCT_PARAMS AS OBJECT (ELEMENTO VARCHAR2(2000), SEPARATORE VARCHAR2(10));

CREATE TYPE T_LISTA_ELEMENTI AS TABLE OF VARCHAR2(2000);

CREATE TYPE T_LISTAGG_DISTINCT AS OBJECT (

    LISTA_ELEMENTI T_LISTA_ELEMENTI,
        SEPARATORE VARCHAR2(10),

    STATIC FUNCTION ODCIAGGREGATEINITIALIZE(SCTX  IN OUT            T_LISTAGG_DISTINCT) 
                    RETURN NUMBER,

    MEMBER FUNCTION ODCIAGGREGATEITERATE   (SELF  IN OUT            T_LISTAGG_DISTINCT, 
                                            VALUE IN                    LISTAGG_DISTINCT_PARAMS ) 
                    RETURN NUMBER,

    MEMBER FUNCTION ODCIAGGREGATETERMINATE (SELF         IN     T_LISTAGG_DISTINCT,
                                            RETURN_VALUE OUT    VARCHAR2, 
                                            FLAGS        IN     NUMBER      )
                    RETURN NUMBER,

    MEMBER FUNCTION ODCIAGGREGATEMERGE       (SELF               IN OUT T_LISTAGG_DISTINCT,
                                                                                        CTX2                 IN         T_LISTAGG_DISTINCT    )
                    RETURN NUMBER
);

CREATE OR REPLACE TYPE BODY T_LISTAGG_DISTINCT IS 

    STATIC FUNCTION ODCIAGGREGATEINITIALIZE(SCTX IN OUT T_LISTAGG_DISTINCT) RETURN NUMBER IS 
    BEGIN
                SCTX := T_LISTAGG_DISTINCT(T_LISTA_ELEMENTI() , ',');
        RETURN ODCICONST.SUCCESS;
    END;

    MEMBER FUNCTION ODCIAGGREGATEITERATE(SELF IN OUT T_LISTAGG_DISTINCT, VALUE IN LISTAGG_DISTINCT_PARAMS) RETURN NUMBER IS
    BEGIN

                IF VALUE.ELEMENTO IS NOT NULL THEN
                        SELF.LISTA_ELEMENTI.EXTEND;
                        SELF.LISTA_ELEMENTI(SELF.LISTA_ELEMENTI.LAST) := TO_CHAR(VALUE.ELEMENTO);
                        SELF.LISTA_ELEMENTI:= SELF.LISTA_ELEMENTI MULTISET UNION DISTINCT SELF.LISTA_ELEMENTI;
                        SELF.SEPARATORE := VALUE.SEPARATORE;
                END IF;
        RETURN ODCICONST.SUCCESS;
    END;

    MEMBER FUNCTION ODCIAGGREGATETERMINATE(SELF IN T_LISTAGG_DISTINCT, RETURN_VALUE OUT VARCHAR2, FLAGS IN NUMBER) RETURN NUMBER IS
      STRINGA_OUTPUT            CLOB:='';
            LISTA_OUTPUT                T_LISTA_ELEMENTI;
            TERMINATORE                 VARCHAR2(3):='...';
            LUNGHEZZA_MAX           NUMBER:=4000;
    BEGIN

                IF SELF.LISTA_ELEMENTI.EXISTS(1) THEN -- se esiste almeno un elemento nella lista

                        -- inizializza una nuova lista di appoggio
                        LISTA_OUTPUT := T_LISTA_ELEMENTI();

                        -- riversamento dei soli elementi in DISTINCT
                        LISTA_OUTPUT := SELF.LISTA_ELEMENTI MULTISET UNION DISTINCT SELF.LISTA_ELEMENTI;

                        -- ordinamento degli elementi
                        SELECT CAST(MULTISET(SELECT * FROM TABLE(LISTA_OUTPUT) ORDER BY 1 ) AS T_LISTA_ELEMENTI ) INTO LISTA_OUTPUT FROM DUAL;

                        -- concatenazione in una stringa                        
                        FOR I IN LISTA_OUTPUT.FIRST .. LISTA_OUTPUT.LAST - 1
                        LOOP
                            STRINGA_OUTPUT := STRINGA_OUTPUT || LISTA_OUTPUT(I) || SELF.SEPARATORE;
                        END LOOP;
                        STRINGA_OUTPUT := STRINGA_OUTPUT || LISTA_OUTPUT(LISTA_OUTPUT.LAST);

                        -- se la stringa supera la dimensione massima impostata, tronca e termina con un terminatore
                        IF LENGTH(STRINGA_OUTPUT) > LUNGHEZZA_MAX THEN
                                    RETURN_VALUE := SUBSTR(STRINGA_OUTPUT, 0, LUNGHEZZA_MAX - LENGTH(TERMINATORE)) || TERMINATORE;
                        ELSE
                                    RETURN_VALUE:=STRINGA_OUTPUT;
                        END IF;

                ELSE -- se non esiste nessun elemento, restituisci NULL

                        RETURN_VALUE := NULL;

                END IF;

        RETURN ODCICONST.SUCCESS;
    END;

    MEMBER FUNCTION ODCIAGGREGATEMERGE(SELF IN OUT T_LISTAGG_DISTINCT, CTX2 IN T_LISTAGG_DISTINCT) RETURN NUMBER IS
    BEGIN
        RETURN ODCICONST.SUCCESS;
    END;

END; -- fine corpo

CREATE
FUNCTION LISTAGG_DISTINCT (INPUT LISTAGG_DISTINCT_PARAMS) RETURN VARCHAR2
    PARALLEL_ENABLE AGGREGATE USING T_LISTAGG_DISTINCT;

// Example
SELECT LISTAGG_DISTINCT(LISTAGG_DISTINCT_PARAMS(OWNER, ', ')) AS LISTA_OWNER
FROM SYS.ALL_OBJECTS;

I'm sorry, but in some case (for a very big set), Oracle could return this error:

Object or Collection value was too large. The size of the value
might have exceeded 30k in a SORT context, or the size might be
too big for available memory.

but I think this is a good point of start ;)

Answer 6

Do you mean something like this:

select listagg(the_column, ',') within group (order by the_column)
from (
   select distinct the_column 
   from the_table
) t

If you need more columns, something like this might be what you are looking for:

select col1, listagg(col2, ',') within group (order by col2)
from (
  select col1, 
         col2,
         row_number() over (partition by col1, col2 order by col1) as rn
  from foo
  order by col1,col2
)
where rn = 1
group by col1;