Crosstab function in Postgres returning a one row output when I expect multiple rows

Question

I currently have a table m of the following format:

id    scenario    period    ct
2        1          1       1  
2        1          2       1         


        

Answer 1

This query generates your desired output:

SELECT id, scenario, period, p1, p2, p3, p4  -- all except aux column rn
FROM   crosstab(
  'SELECT row_number() OVER (ORDER BY id, scenario, period)::int AS rn
        , id, scenario, period, period, ct
   FROM   m
   ORDER  BY 1'
, 'VALUES (1), (2), (3), (4)'
   ) AS (rn int, id int, scenario int, period int, p1 int, p2 int, p3 int, p4 int);

Two special difficulties:

1. You don't have a single unique column for the row_name yet. I use row_number() to generate the surrogate key: rn. I removed it from the outer SELECT to match your desired result.
   The way you tried it, id is taken to be the row_name and all input rows are aggregated into a single output row.

2. You want additional columns (scenario and period) in the result, which must come after the row_name and before the category. You must list period twice to get the original column additionally - redundant as though it may seem.

Basics:

• PostgreSQL Crosstab Query

Related to this particular case:

• Pivot on Multiple Columns using Tablefunc

---

Typically, you would have a query like this:

SELECT id, scenario, p1, p2, p3, p4  -- all except aux column rn
FROM   crosstab(
  'SELECT rank() OVER (ORDER BY id, scenario)::int AS rn
        , id, scenario, period, ct
   FROM   m
   ORDER  BY 1'
, 'VALUES (1), (2), (3), (4)'
   ) AS (rn int, id int, scenario int, p1 int, p2 int, p3 int, p4 int);

With an output like this:

id   scenario   p1   p2   p3   p4
2    1          1    1    1    1
2    2          1    1    1    1
2    3          1    1    1    1

Note the use of rank() instead of row_number() to group same combinations of (id, scenario) together.
The result makes more sense if counts are not all 1.