This topic demonstrates how to collect and spread values from multiple rows under a single column, like those in VAL below
GROUP_KEY VAL ---------- ---------- G8 a G8 b G8 c G8 d G8 e G8 f G8 g G8 h
across multiple columns, rows, and groups like this.
GROUP_KEY COL_0 COL_1 COL_2 ---------- ---------- ---------- ---------- G8 a b c G8 d e f G8 g h
Unlike prior topics in this section here we will deal with a scenario where there is no PIVOT_KEY column in the base table. Values will be distributed evenly from left to right, then top to bottom across an arbitrary number of columns.
11g PIVOT Solution
break on group_key duplicates skip 1
column group_key null '(null)'
column val null '(null)'
select
group_key ,
col_0 ,
col_1 ,
col_2
from
( select
group_key,
mod( row_number() over ( partition by group_key order by val ) - 1, 3 )
as pivot_key ,
trunc
( ( row_number() over ( partition by group_key order by val ) - 1 ) / 3 )
as row_num ,
val
from t5
) t5a pivot
( max(val)
for pivot_key in
(
0 as col_0 ,
1 as col_1 ,
2 as col_2
)
)
order by group_key, row_num ;
GROUP_KEY COL_0 COL_1 COL_2 ---------- ---------- ---------- ---------- G0 G1 a G2 a b G3 a b c G4 a b c G4 d G5 a b c G5 d e G6 a b c G6 d e f G7 a b c G7 d e f G7 g G8 a b c G8 d e f G8 g h
To spread values across 5 columns instead of 3 the query would look like this.
select
group_key ,
col_0 ,
col_1 ,
col_2 ,
col_3 , -- add this line
col_4 -- add this line
from
( select
group_key,
mod( row_number() over ( partition by group_key order by val ) - 1, 5 ) -- use "5" here
as pivot_key ,
trunc
( ( row_number() over ( partition by group_key order by val ) - 1 ) / 5 ) -- use "5" here
as row_num ,
val
from t5
) t5a pivot
( max(val)
for pivot_key in
(
0 as col_0 ,
1 as col_1 ,
2 as col_2 ,
3 as col_3 , -- add this line
4 as col_4 -- add this line
)
)
order by group_key, row_num ;
GROUP_KEY COL_0 COL_1 COL_2 COL_3 COL_4 ---------- ---------- ---------- ---------- ---------- ---------- G0 G1 a G2 a b G3 a b c G4 a b c d G5 a b c d e G6 a b c d e G6 f G7 a b c d e G7 f g G8 a b c d e G8 f g h
How it Works
The solution works basically the same as the one presented in Transposing M:1. The only differences are with the inline view, which produces values like these.
break on pivot_key duplicates skip 1
select
group_key,
mod( row_number() over ( partition by group_key order by val ) - 1, 3 )
as pivot_key ,
trunc
( ( row_number() over ( partition by group_key order by val ) - 1 ) / 3 )
as row_num ,
val
from t5
where group_key = 'G8'
order by group_key, pivot_key, row_num ;
GROUP_KEY PIVOT_KEY ROW_NUM VAL ---------- ---------- ---------- ---------- G8 0 0 a G8 0 1 d G8 0 2 g G8 1 0 b G8 1 1 e G8 1 2 h G8 2 0 c G8 2 1 f
The key feature in the inline view is the PIVOT_KEY column created using MOD( ROW_NUMBER() ....
This column contains an auto generated column number for each VAL value.
This allows us to later place the VAL value in the appropriate column using the PIVOT ... IN clause.
Earlier Versions
For those who are not working with Oracle 11g the following query works in Oracle versions from 8i onward.
break on group_key duplicates skip 1
select
group_key ,
max( decode( pivot_key, 0, val, null ) ) col_0 ,
max( decode( pivot_key, 1, val, null ) ) col_1 ,
max( decode( pivot_key, 2, val, null ) ) col_2
from
( select
group_key,
mod( row_number() over ( partition by group_key order by val ) - 1, 3 )
as pivot_key ,
trunc
( ( row_number() over ( partition by group_key order by val ) - 1 ) / 3 )
as row_num ,
val
from t5
) t5a
group by group_key, row_num
order by group_key, row_num ;
GROUP_KEY COL_0 COL_1 COL_2 ---------- ---------- ---------- ---------- G0 G1 a G2 a b G3 a b c G4 a b c G4 d G5 a b c G5 d e G6 a b c G6 d e f G7 a b c G7 d e f G7 g G8 a b c G8 d e f G8 g h