Pivot a table using a stored procedure - stored-procedures

Question

So I'm trying to pivot a table using a stored procedure in Redshift. The issue is that the result set is dynamic. That means that we'd need to be able to dynamically pivot the table below. This is what I am trying to pivot:
| object_uid | field | value |
|------------|----------|----------|
| post:1 | field_1_a| test |
| post:2 | field_2_a| turtle |
| post:2 | field_2_b| frog |
| post:3 | field_3_a| mountain |
| ...... | ..... | ...... |
|------------|----------|----------|
This would be pivoted into the following:
| object_uid | field_1_a| field_2_a| field_2_b| field_3_a|
|------------|----------|----------|----------|----------|
| post:1 | test | | | |
| post:2 | | turtle | frog | |
| post:3 | | | | mountain |
| ....... | ..... | ....... | ....... | ....... |
|------------|----------|----------|----------|----------|
Essentially I am trying to construct a chained string of column names (the field_* columns) via SELECT LISTAGG statement in the subquery, and trying to interpolate that statement's output in the CREATE TABLE sql statement. Then once the CREATE TABLE sql statement is constructed, the sql statement gets executed via the EXECUTE command.
However, This is not behaving as expected. I am a relative newcomer to Redshift, so I apologize in advance if this is a terrible way to go about pivoting a table from a tall one to a wide one. This is the code that I have so far:
CREATE OR REPLACE PROCEDURE my_proc (
tmp_name INOUT varchar(256)
) AS $$
DECLARE
sql VARCHAR(MAX) := '';
BEGIN
WITH pivot_output AS (
SELECT LISTAGG(temp.output, ', ') WITHIN GROUP (ORDER BY temp.output) AS metadata FROM
(
SELECT DISTINCT
'MAX(IF(cm.metadata = ''' || metadata || ''',cm.field_value,NULL)) AS ' || QUOTE_LITERAL(metadata)
AS output
FROM "content_metadata" cm
WHERE cm."source_uid_type" = 'post'
) AS temp
);
sql = 'CREATE TABLE ' || tmp_name || ' AS SELECT cm.object_uid, ' || pivot_output.metadata || ' FROM content_metadata cm GROUP BY cm.object_uid';
EXECUTE sql;
END;
$$ LANGUAGE plpgsql;
CALL my_proc ('output_table');
I get the following error when trying to execute the above:
The database reported a syntax error: Amazon Invalid operation: syntax error at or near "$1";
A little bit stumped by the error. Does anyone have any clues / suggestions?

Answer 1

I added a working example of emulating the PIVOT FOR syntax in our GitHub repo "Amazon Redshift Utils". https://github.com/awslabs/amazon-redshift-utils/blob/master/src/StoredProcedures/sp_pivot_for.sql
I hope this is useful for you. Let me know if you have any issues with it.