Question

A_first B_first C_first  A_second  B_second  C_second A_third  B_third  C_third
638      450     188       638       439       187      546      256      789

I have a table like that with only one row. How can I pivot this row so I get A B C as columns and a row for each _first, _second, _third?

I’m trying to wrap my head around the pivot/unpivot command in sql

A          B      C     
638      450     188     
638      439     187     
546      256     789

Using SQL Server 2008 so I have the pivot/unpivot command. More specifically SSMS tells me the database I’m connected to is 9.0 SP3

Answer 1

You did not specify what RDBMS you are using so here are two solutions. If you have an UNPIVOT function and then a PIVOT:

select *
from
(
  select value, left(fields, 1) col,
    substring(fields, 3, len(fields) -2) row
  from yourtable
  unpivot
  (
    value
    for fields in(A_first, B_first, C_first,
                  A_second, B_second, C_second,
                  A_third, B_third, C_third)
  ) unpiv
) src
pivot
(
  max(value)
  for col in ([A], [B], [C])
) piv

See SQL Fiddle with Demo

If you do not have an UNPIVOT and PIVOT function, then you can use UNION ALL:

select
  max(case when col = 'A' then value end) A,
  max(case when col = 'B' then value end) B,
  max(case when col = 'C' then value end) C
from
(
  select value, left(fields, 1) col, substring(fields, 3, len(fields) -2) row
  from
  (
    select A_first value, 'A_first' fields
    from yourtable
    union all
    select B_first value, 'B_first' fields
    from yourtable
    union all
    select C_first value, 'C_first' fields
    from yourtable
    union all
    select A_second value, 'A_second' fields
    from yourtable
    union all
    select B_second value, 'B_second' fields
    from yourtable
    union all
    select C_second value, 'C_second' fields
    from yourtable
    union all
    select A_third value, 'A_third' fields
    from yourtable
    union all
    select B_third value, 'B_third' fields
    from yourtable
    union all
    select C_third value, 'C_third' fields
    from yourtable
  ) unpiv
) src
group by row

See SQL Fiddle with Demo

Both will produce the same result:

|   A |   B |   C |
-------------------
| 638 | 450 | 188 |
| 638 | 439 | 187 |
| 546 | 256 | 789 |

Note, in order to perform this task the data types must be the same on each column. If they are not then you will need to convert the data first prior to the UNPIVOT.

This answer makes a few assumptions:

1. That the existing column will always start with the new column value (A,
   B or C, etc)
2. That the existing column will always start with two characters
   before the row number.