>
> The only issue I find with your SQL is it relies on "(a.attnum = x.conkey[1] or
> a.attnum = x.conkey[2])" which assumes
there is two columns forming the primary key of a table. Perhaps, I should explain
what I'm trying to achieve with this
SQL.
>
Not exactly. Look at this
CREATE TABLE films (
code CHAR(5),
title VARCHAR(40),
did DECIMAL(3),
date_prod DATE,
kind VARCHAR(10),
len INTERVAL HOUR TO MINUTE,
CONSTRAINT code_title PRIMARY KEY(code,title)
);
CREATE TABLE films_02 (
code CHARACTER(5) CONSTRAINT firstkey PRIMARY KEY,
title CHARACTER VARYING(40) NOT NULL,
did DECIMAL(3) NOT NULL,
date_prod DATE,
kind CHAR(10),
len INTERVAL HOUR TO MINUTE
);
SELECT
upper(u.usename) AS TBL_OWNER, upper(c.relname) AS TBL_NAME,
upper(a.attname) AS COL_NAME, a.atttypid AS COL_TYPE,
int4larger(a.attlen, a.atttypmod - 4) AS COL_LENGTH,
CASE WHEN a.attnotnull=true THEN 0 ELSE 1 END AS COL_NULL, a.attnum AS COL_SEQ,
CASE WHEN EXISTS(SELECT adsrc FROM pg_attrdef d
WHERE d.adrelid = a.attrelid and d.adnum = a.attnum) THEN
1
ELSE
0
END AS COL_DEFAULT
from pg_attribute a, pg_constraint x,
pg_class c left join pg_user u on (u.usesysid = c.relowner)
where c.oid = a.attrelid and not (c.relname ~* 'pg_') and
c.relkind = 'r' and a.attnum > 0 and
c.relfilenode=x.conrelid and x.contype='p' and c.relname like 'film%' and
(a.attnum = x.conkey[1] or a.attnum = x.conkey[2]) ;
tbl_owner | tbl_name | col_name | col_type | col_length | col_null | col_seq |
col_default
-----------+----------+----------+----------+------------+----------+---------+-------------
CH | FILMS | CODE | 1042 | 5 | 0 | 1 |
0
CH | FILMS | TITLE | 1043 | 40 | 0 | 2 |
0
CH | FILMS_02 | CODE | 1042 | 5 | 0 | 1 |
0
(3 rows)
My supposition was it's no sign of good design to form primary keys from more
than two columns.
As you can see primary keys on a single column will be listed as well.
So, if you want to deal with wider keys,
you can add "or a.attnum = x.conkey[3]" etc.
> I'm writing an abstraction layer (abit like persistent but less ambitious). I hope
> to make this C++ layer generic for
any RDBMS as long as the RDMS allow discovery of a table's columns name and the
columns' data type. The RDBMS should als
o expose the field(s) used to form the primary key of a table. This will free the
programmer from coding the class data
member to correspond to the underlying table's fields (automate those tedious tasks of
mapping OO classes to database ta
bles).
Good Idea. But I've found it's always a PITA to question system tables,
no matter which RDBMS you're on.
> I'm using libpqxx for postgreSQL, I had thought of a hack which is to strip the
> {1,2..} string returned by the conkey
of pg_constraint to get the column numbers. It's not pretty as I have to execute at
least two queries but it should be w
orkable.
Anyway.
Regards, Christoph
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
