[
https://issues.apache.org/jira/browse/DERBY-6362?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13789244#comment-13789244
]
Rick Hillegas commented on DERBY-6362:
--------------------------------------
I have exchanged email with some other members of the SQL committee. They have
pointed me at the portions of the SQL Standard which discuss the SQL-path. It
sheds some light on this discussion.
The SQL-path is introduced by part 2, section 4.29 (SQL-paths). Fully evolved,
a SQL-path is a search order for schema resolution. It's akin to the search
order in a Unix path or Java classpath. The CURRENT_PATH variable (implemented
in some databases) shows the value of the SQL-path...
values CURRENT_PATH
...much as the following Unix commands show the value of the shell's path and
classpath variables:
echo $PATH
echo $CLASSPATH
According to part 2, section 4.29 (SQL-paths), the SQL-path is used to resolve
routine and type references which lack a schema qualifier and which appear in
"preparable statements". These statements, in turn, include the standalone DDL
statements like CREATE/ALTER TABLE. So SQL-path should be used to resolve
unqualified routine/type references in CHECK constraints and generation clauses.
The value of SQL-path is implementation-defined. The Standard lets you change
the value of SQL-path via a SET PATH command, which Derby does not implement.
SET PATH is independent of the SET SCHEMA command which Derby does implement.
Note that the SET SCHEMA command (part 2, section 19.6) changes how unqualified
names are resolved in "preparable statements". However, that section does not
claim that SET SCHEMA has any affect on SQL-path.
The Derby code never speaks about SQL-path. This is a part of the Standard
which Derby ignores. So Derby's inconsistent treatment of SQL-path is not
surprising. That said, now that we are aware of SQL-path, it would be good to
make Derby's behavior consistent.
For most statements, Derby behaves as though CURRENT_PATH is a list with one
value in it (CURRENT_SCHEMA). So, unqualified routine references in DML
statements are resolved to CURRENT_SCHEMA. The same is true for unqualified
routine references in generation clauses and unqualified type references in
CREATE TABLE statements.
If we treated CHECK constraints the same way, then we would have behavior (C)
described above. This seems reasonable to me. That is, it makes sense to me
that we should formalize our understand of SQL-path and apply it consistently
going forward. Derby's implementation of SQL-path would be simple, viz.,
CURRENT_PATH = CURRENT_SCHEMA.
The following script shows the current behavior of CURRENT_PATH:
---------------------------------------------
connect 'jdbc:derby:memory:db;create=true';
create schema neutralSchema;
create schema ddlSessionSchema;
set schema ddlSessionSchema;
create type myType external name 'java.util.HashMap' language java;
create function myCheckFunc( doubleVal double ) returns double
language java parameter style java deterministic no sql
external name 'java.lang.Math.sin';
create function myGenColFunc( doubleVal double ) returns double
language java parameter style java deterministic no sql
external name 'java.lang.Math.abs';
-- --------------------------------
create schema tableSchema;
set schema tableSchema;
create type myType external name 'java.util.ArrayList' language java;
create function myCheckFunc( doubleVal double ) returns double
language java parameter style java deterministic no sql
external name 'java.lang.Math.cos';
create function myGenColFunc( doubleVal double ) returns double
language java parameter style java deterministic no sql
external name 'java.lang.Math.exp';
-- --------------------------------
set schema ddlSessionSchema;
create table tableSchema.t
(
a myType,
b double check ( myCheckFunc( b ) > 0 ),
c generated always as ( myGenColFunc( b ) )
);
-- --------------------------------
set schema neutralSchema;
values
(
ddlSessionSchema.myCheckFunc( -1 ), ddlSessionSchema.myGenColFunc( -1 )
);
values
(
tableSchema.myCheckFunc( -1 ), tableSchema.myGenColFunc( -1 )
);
-- fails because myType resolves to ddlSessionSchema.myType
insert into tableSchema.t( a ) values ( cast (null as tableSchema.myType) );
--
-- myCheckFunc resolves to tableSchema.myCheckFunc
--
-- myGenColFunc resolves to ddlSessionSchema.myGenColFunc
--
insert into tableSchema.t( b ) values ( -1 );
select * from tableSchema.t;
-- --------------------------------
set schema ddlSessionSchema;
-- behaves as though CURRENT_PATH = ddlSessionSchema
values
(
myCheckFunc( -1 ), myGenColFunc( -1 )
);
set schema tableSchema;
-- behaves as though CURRENT_PATH = tableSchema
values
(
myCheckFunc( -1 ), myGenColFunc( -1 )
);
> CHECK constraint uses wrong schema for unqualified routine invocations
> ----------------------------------------------------------------------
>
> Key: DERBY-6362
> URL: https://issues.apache.org/jira/browse/DERBY-6362
> Project: Derby
> Issue Type: Bug
> Components: SQL
> Affects Versions: 10.10.1.1
> Reporter: Knut Anders Hatlen
>
> DERBY-3944 fixed the problem with CHECK constraints invoking different
> routines depending on who performed the triggering INSERT or UPDATE statement.
> The discussion leading up to DERBY-3944 can be found here:
> http://mail-archives.apache.org/mod_mbox/db-derby-dev/200811.mbox/%[email protected]%3E
> Three alternatives are discussed in the thread:
> A) The schema that holds the CHECK constraint?
> B) The schema that holds the table?
> C) The current schema when the CREATE TABLE statement was issued?
> The conclusion in the thread was that option C was the correct one. However,
> what was implemented, was option B.
> I cannot find any information in DERBY-3944 about why option B ended up being
> chosen, so I assume that it was unintended.
> Here's an ij script that shows how the CHECK constraint tries to invoke the
> TO_HEX function in the schema of the target table (S2) instead of the schema
> that was the current schema at the time of CREATE TABLE:
> ij version 10.10
> ij> connect 'jdbc:derby:memory:db;create=true';
> ij> create schema s1;
> 0 rows inserted/updated/deleted
> ij> create schema s2;
> 0 rows inserted/updated/deleted
> ij> create function s1.to_hex(i int) returns char(4) language java parameter
> style java external name 'java.lang.Integer.toHexString' no sql;
> 0 rows inserted/updated/deleted
> ij> set schema s1;
> 0 rows inserted/updated/deleted
> ij> create table s2.t(x int, constraint cc check(to_hex(x) <> '80'));
> 0 rows inserted/updated/deleted
> ij> insert into s2.t values 1;
> ERROR 42Y03: 'TO_HEX' is not recognized as a function or procedure.
> (errorCode = 30000)
> ij> create function s2.to_hex(i int) returns char(4) language java parameter
> style java external name 'java.lang.Integer.toHexString' no sql;
> 0 rows inserted/updated/deleted
> ij> insert into s2.t values 1;
> 1 row inserted/updated/deleted
--
This message was sent by Atlassian JIRA
(v6.1#6144)
