Concatenation of character values by columns with excluding duplicates

Question

Hi all! Assume, that there is a multi-rows table of the following form (for example):
1|A  |   |B  |C  |   |D  |   |
2|   |   |A  |B  |C  |D  |   |
3|   |A  |B  |   |   |C  |D  |
4|   |   |A  |   |B  |C  |D  |
5|   |   |A  |   |B  |   |C  |i.e. there are NULL and NOTNULL values in VARCHAR2( or NVARCHAR2) format, and the first column is a primary key.
I need to do next:
1. Select some set of certain rows, primary keys of those are initialy known, e.g. 1, 3, 5;
2. Parse this table such way: concatenate values by columns in selected rows, if there are few not-null values in one column, it's necessary to separate concatenated values with comma and whitesbase (it's desirable that's word after comma with whitespace is outputted in lowercase). Finally, it's necessary to insert dot and whitespace character after the last value.
3. Exclude duplicate substrings from the outcoming text values.
Thereafter I need to initialize set of local variables with the values, obtained by forementioned parsing (parsing of one column match the corresponding variable), e.g. columns of 1, 3 and 5 rows.
I.e. result in our case looks like:
variable1 = A.
variable2 = A.
variable3 = B, A.
variable4 = C.
variable5 = B.
variable6 = D, C.
variable7 = D, C.
Does the task will be easier, if the number of parsed rows is known (e.g. 3 as in our case)?
Edited by: 942736 on 02.10.2012 17:17

Answer 1

You missed your db-version. Please give the result from
select * from v$version;When you are on 11.2 you can try this (will not work with 10.x):
with yourtable as
select 1 id, 'A' col1, 'N' col2, 'B' col3, 'C' col4 ,'N' col5, 'D' col6 ,'N' col7 from dual union all
select 2,'N','N','A','B','C','D','N' from dual union all
select 3,'N','A','B','N','N','C','D' from dual union all
select 4,'N','N','A','N','B','C','D' from dual union all
select 5,'N','N','A','N','B','N','C' from dual
select
(select listagg(col1,',') within group (order by id)||'.' from (select distinct col1, min(id) over (partition by col1 order by id) id from yourtable where id in (1,3,5))) v1,
(select listagg(col2,',') within group (order by id)||'.' from (select distinct col2, min(id) over (partition by col2 order by id) id from yourtable where id in (1,3,5))) v2,
(select listagg(col3,',') within group (order by id)||'.' from (select distinct col3, min(id) over (partition by col3 order by id) id from yourtable where id in (1,3,5))) v3,
(select listagg(col4,',') within group (order by id)||'.' from (select distinct col4, min(id) over (partition by col4 order by id) id from yourtable where id in (1,3,5))) v4,
(select listagg(col5,',') within group (order by id)||'.' from (select distinct col5, min(id) over (partition by col5 order by id) id from yourtable where id in (1,3,5))) v5,
(select listagg(col6,',') within group (order by id)||'.' from (select distinct col6, min(id) over (partition by col6 order by id) id from yourtable where id in (1,3,5))) v6,
(select listagg(col7,',') within group (order by id)||'.' from (select distinct col7, min(id) over (partition by col7 order by id) id from yourtable where id in (1,3,5))) v7
from dual;Output with 11.2 is:
V1      V2      V3      V4      V5      V6      V7
A,N.     N,A.     B,A.     C,N.     N,B.     D,C,N.     N,D,C.@ranit B: You missed, that the OP said that only rows 1,3 and 5 shall be used.
Edited by: hm on 01.10.2012 23:09