How to make a long table wide without aggreagation? - sql

Question

I have this table of financial transactions..
PersonID | SeqId | FundId | PortfolioDbu | Date
----------------------------------------------------------
456 | 1 | B | 0.1 | 2012-04-03
456 | 1 | F | 0.5 | 2012-04-03
456 | 1 | H | 0.3 | 2012-04-03
456 | 1 | Z | 0.1 | 2012-04-03
8 | 1 | B | 0.5 | 2012-03-23
8 | 1 | A | 0.5 | 2012-03-23
8 | 2 | C | 0.3 | 2011-03-24
8 | 2 | X | 0.3 | 2011-03-24
8 | 2 | F | 0.4 | 2011-03-24
6001 | 1 | J | 0.5 | 2008-01-01
6001 | 1 | R | 0.5 | 2008-01-01
76 | 1 | A | 0.25 | 2010-09-26
76 | 1 | B | 0.25 | 2010-09-26
76 | 1 | C | 0.25 | 2010-09-26
76 | 1 | D | 0.25 | 2010-09-26
321 | 1 | X | 0.2 | 2012-02-21
321 | 1 | Y | 0.2 | 2012-02-21
321 | 1 | U | 0.2 | 2012-02-21
321 | 1 | P | 0.2 | 2012-02-21
321 | 1 | W | 0.2 | 2012-02-21
456 | 2 | Y | 1 | 2012-11-01
which I need to convert to a "wide" format, like so..
Date | PersonId | SeqId | Fund1 | Fund2 | Fund3 | Fund4 | Fund5 | Dbu1 | Dbu2 | Dbu3 | Dbu4 | Dbu5
----------------------------------------------------------------------------------------------------------
2012-04-03 | 456 | 1 | B | F | H | Z | . | 0.1 | 0.5 | 0.3 | 0.1 | .
2012-03-23 | 8 | 1 | B | A | . | . | . | 0.5 | 0.5 | . | . | .
2012-03-24 | 8 | 2 | C | X | F | . | . | 0.3 | 0.3 | 0.4 | . | .
2008-01-01 | 6001 | 1 | J | R | . | . | . | 0.5 | 0.5 | . | . | .
2010-09-26 | 76 | 1 | A | B | C | D | . | 0.25 | 0.25 | 0.25 | 0.25 | .
2010-02-21 | 321 | 1 | X | Y | U | P | W | 0.2 | 0.2 | 0.2 | 0.2 | 0.2
2012-11-01 | 456 | 2 | Y | . | . | . | . | 1 | . | . | . | .
Is this possible even though I don't want to aggregate the data in any way?
SQL Fiddle

Answer 1

I'm not real good a PIVOT tables, but you can use the following alternative CASE statement pattern to get the output you're looking for:
WITH T AS (
SELECT
personid,
seqid,
row_number() over (partition BY personid,seqid ORDER BY FundId) AS ROW,
FundId,
portfoliodbu,
date
FROM
transactions
)
SELECT
date,
personid,
seqid,
max(CASE WHEN ROW=1 THEN fundid END) AS fund1,
max(CASE WHEN ROW=2 THEN fundid END) AS fund2,
max(CASE WHEN ROW=3 THEN fundid END) AS fund3,
max(CASE WHEN ROW=4 THEN fundid END) AS fund4,
max(CASE WHEN ROW=5 THEN fundid END) AS fund5,
max(CASE WHEN ROW=1 THEN portfoliodbu END) AS dbu1,
max(CASE WHEN ROW=2 THEN portfoliodbu END) AS dbu2,
max(CASE WHEN ROW=3 THEN portfoliodbu END) AS dbu3,
max(CASE WHEN ROW=4 THEN portfoliodbu END) AS dbu4,
max(CASE WHEN ROW=5 THEN portfoliodbu END) AS dbu5
FROM
T
GROUP BY
date,personid,seqid
Demo: SQL Fiddle
Results:
| DATE | PERSONID | SEQID | FUND1 | FUND2 | FUND3 | FUND4 | FUND5 | DBU1 | DBU2 | DBU3 | DBU4 | DBU5 |
----------------------------------------------------------------------------------------------------------------------------------------------
| January, 01 2008 00:00:00+0000 | 6001 | 1 | J | R | (null) | (null) | (null) | 0.5 | 0.5 | (null) | (null) | (null) |
| September, 26 2010 00:00:00+0000 | 76 | 1 | A | B | C | D | (null) | 0.25 | 0.25 | 0.25 | 0.25 | (null) |
| March, 24 2011 00:00:00+0000 | 8 | 2 | C | F | X | (null) | (null) | 0.3 | 0.4 | 0.3 | (null) | (null) |
| February, 21 2012 00:00:00+0000 | 321 | 1 | P | U | W | X | Y | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
| March, 23 2012 00:00:00+0000 | 8 | 1 | A | B | (null) | (null) | (null) | 0.5 | 0.5 | (null) | (null) | (null) |
| April, 03 2012 00:00:00+0000 | 456 | 1 | B | F | H | Z | (null) | 0.1 | 0.5 | 0.3 | 0.1 | (null) |
| November, 01 2012 00:00:00+0000 | 456 | 2 | Y | (null) | (null) | (null) | (null) | 1 | (null) | (null) | (null) | (null) |