Thread: Table Constraint CHECK(SELECT()) alternative?

Hello,
I have seen in another thread that sub-queries in a CHECK constraint have implementation ramifications that make them
awkwardto implement and support. OK, fair enough, c'est la vie.
 

ERROR:  cannot use subquery in check constraint

is the result. I have a model which seems to BEG for just such a feature though. What alternatives are there to the use
ofCHECK(SELECT()) in this model?
 
I am implementing a basic tree of nodes. There are leafs and directories... node types. Each node in a directory has a
uniquename, and the node type ids point to a table with node type names. Each node points to a valid node (its parent).
Thereis more... but now to my question. I wish to constrain parent nodes to only be directories type nodes. Leafs
cannotbe parents.
 
                                       Table "public.raw_pnfs_nodes"    Column     |         Type          |
                 Modifiers
 
----------------+-----------------------+---------------------------------------------------------------------node_id
    | integer               | not null default nextval('public.raw_pnfs_nodes_node_id_seq'::text)parent_node_id |
integer              | not nullnode_name      | character varying(80) | not nullnode_type_id   | smallint
|not null
 
Indexes:   "raw_pnfs_nodes_pkey" primary key, btree (node_id)   "raw_pnfs_nodes_node_name_key" unique, btree
(node_name,parent_node_id)
 
Foreign-key constraints:   "parent_refersto_node" FOREIGN KEY (parent_node_id) REFERENCES raw_pnfs_nodes(node_id) ON
UPDATERESTRICT ON DELETE RESTRICT   "raw_pnfs_nodes" FOREIGN KEY (node_type_id) REFERENCES node_types(node_type_id) ON
UPDATERESTRICT ON DELETE RESTRICT
 


What I WANT to write is the table constraint:
 CONSTRAINT parents_are_dirs_check CHECK( (SELECT A.node_type_id FROM raw_pnfs_nodes A, raw_pnfs_nodes B WHERE
A.node_id= B.parent_node_id) = 1)
 

where "1" is the id for directory node types (ok, this is not polished yet) Hopefully this illustrates the idea. The
nodetype of a node that is a parent to another node must be the directory node type. Is there another way to express
thisin SQL? I would like to avoid putting this into the business logic layer.
 

Thanks much,
RDK

On Sat, 3 Sep 2005, Robert D. Kennedy wrote:

>     I have seen in another thread that sub-queries in a CHECK
> constraint have implementation ramifications that make them awkward to
> implement and support. OK, fair enough, c'est la vie.
>
> ERROR:  cannot use subquery in check constraint
>
> is the result. I have a model which seems to BEG for just such a feature
> though. What alternatives are there to the use of CHECK(SELECT()) in
> this model?

The simplest thing is to place the select into a function and use a check
constraint on that (with appropriate row locks).  However, that won't
prevent removal or change of a referenced row.

You can use triggers to prevent that (in your case to prevent removal or
change of id of a referenced node or the change of type of a referenced
mode away from directory).  At that point, you might be just as well off
using a trigger to do the initial check as well.

You can create a function and call it from a trigger on that column insert
event
Thanks
Dinesh Pandey

-----Original Message-----
From: pgsql-sql-owner@postgresql.org [mailto:pgsql-sql-owner@postgresql.org]
On Behalf Of Robert D. Kennedy
Sent: Sunday, September 04, 2005 12:20 AM
To: pgsql-sql@postgresql.org
Cc: rdkennedy@acm.org
Subject: [SQL] Table Constraint CHECK(SELECT()) alternative?

Hello,
I have seen in another thread that sub-queries in a CHECK constraint
have implementation ramifications that make them awkward to implement and
support. OK, fair enough, c'est la vie.

ERROR:  cannot use subquery in check constraint

is the result. I have a model which seems to BEG for just such a feature
though. What alternatives are there to the use of CHECK(SELECT()) in this
model?
I am implementing a basic tree of nodes. There are leafs and
directories... node types. Each node in a directory has a unique name, and
the node type ids point to a table with node type names. Each node points to
a valid node (its parent). There is more... but now to my question. I wish
to constrain parent nodes to only be directories type nodes. Leafs cannot be
parents.
                                       Table "public.raw_pnfs_nodes"    Column     |         Type          |
Modifiers
----------------+-----------------------+-----------------------------------
----------------------------------node_id        | integer               | not null default
nextval('public.raw_pnfs_nodes_node_id_seq'::text)parent_node_id | integer               | not nullnode_name      |
charactervarying(80) | not nullnode_type_id   | smallint              | not null
 
Indexes:   "raw_pnfs_nodes_pkey" primary key, btree (node_id)   "raw_pnfs_nodes_node_name_key" unique, btree
(node_name,parent_node_id)
 
Foreign-key constraints:   "parent_refersto_node" FOREIGN KEY (parent_node_id) REFERENCES
raw_pnfs_nodes(node_id) ON UPDATE RESTRICT ON DELETE RESTRICT   "raw_pnfs_nodes" FOREIGN KEY (node_type_id) REFERENCES
node_types(node_type_id) ON UPDATE RESTRICT ON DELETE RESTRICT


What I WANT to write is the table constraint:
 CONSTRAINT parents_are_dirs_check CHECK( (SELECT A.node_type_id FROM
raw_pnfs_nodes A, raw_pnfs_nodes B WHERE A.node_id = B.parent_node_id) = 1)

where "1" is the id for directory node types (ok, this is not polished yet)
Hopefully this illustrates the idea. The node type of a node that is a
parent to another node must be the directory node type. Is there another way
to express this in SQL? I would like to avoid putting this into the business
logic layer.

Thanks much,
RDK