How to get column attributes query from table name using PostgreSQL?

Built on the answer by @Chris:

SELECT a.attnum
      ,a.attname                            AS name
      ,format_type(a.atttypid, a.atttypmod) AS typ
      ,a.attnotnull                         AS notnull
      ,coalesce(p.indisprimary, FALSE)      AS primary_key
      ,f.adsrc                              AS default_val
      ,d.description                        AS col_comment
FROM   pg_attribute    a 
LEFT   JOIN pg_index   p ON p.indrelid = a.attrelid AND a.attnum = ANY(p.indkey)
LEFT   JOIN pg_description d ON d.objoid  = a.attrelid AND d.objsubid = a.attnum
LEFT   JOIN pg_attrdef f ON f.adrelid = a.attrelid  AND f.adnum = a.attnum
WHERE  a.attnum > 0
AND    NOT a.attisdropped
AND    a.attrelid = 'schema.tbl'::regclass  -- table may be schema-qualified
ORDER  BY a.attnum;

But:

Table names are not unique in a database and hence also not in the system catalog. You may have to schema-qualify the name.
Use a.attrelid = 'tbl'::regclass as condition. This way you can pass myschema.mytbl as name and disambiguate. Then there is no need to join to pg_class at all in this case.
Also, visibility is checked automatically for regclass and there is no need for pg_table_is_visible().

A primary key can span multiple columns. I take care of this by joining to pg_index on a.attnum = ANY(p.indkey).
indkey is of type int2vecor, which is a special case of int2[], only used in the catalogs.

I find psql -E helpful for this class of problems.

Compatibility

A specialized query like this might break after a major version update. Postgres does not guarantee that catalog tables remain stable. It is extremely unlikely that basic elements change, but the more complex and specialized your query gets, the bigger the chance. You could use the information schema instead, which is standardized, but also comparatively slow.

Here's query against the system catalog that should fetch everything you need (with a bonus primary-key field thrown in for free).

SELECT DISTINCT
    a.attnum as num,
    a.attname as name,
    format_type(a.atttypid, a.atttypmod) as typ,
    a.attnotnull as notnull, 
    com.description as comment,
    coalesce(i.indisprimary,false) as primary_key,
    def.adsrc as default
FROM pg_attribute a 
JOIN pg_class pgc ON pgc.oid = a.attrelid
LEFT JOIN pg_index i ON 
    (pgc.oid = i.indrelid AND i.indkey[0] = a.attnum)
LEFT JOIN pg_description com on 
    (pgc.oid = com.objoid AND a.attnum = com.objsubid)
LEFT JOIN pg_attrdef def ON 
    (a.attrelid = def.adrelid AND a.attnum = def.adnum)
WHERE a.attnum > 0 AND pgc.oid = a.attrelid
AND pg_table_is_visible(pgc.oid)
AND NOT a.attisdropped
AND pgc.relname = 'TABLE_NAME'  -- Your table name here
ORDER BY a.attnum;

Which would return results like:

 num |    name     |             typ             | notnull |       comment       | primary_key 
-----+-------------+-----------------------------+---------+---------------------+-------------
   1 | id          | integer                     | t       | a primary key thing | t
   2 | ref         | text                        | f       |                     | f
   3 | created     | timestamp without time zone | t       |                     | f
   4 | modified    | timestamp without time zone | t       |                     | f
   5 | name        | text                        | t       |                     | f

• num: The column number
• name: The column name
• typ: the data type
• notnull: Is the column defined as NOT NULL
• comment: Any COMMENT defined for the column
• primary_key: Is the column defined as PRIMARY KEY
• default: The command used for the default value