Vous n'êtes pas identifié(e).

#1 Re : Général » audit table with integer column » 25/07/2016 17:52:05

Finaly, i'm not sure the problem was do to CONCAT ignoring NULL argument but to the fact that in the CASE WHEN i was returning either an integer 'NEW."categorySignId" or varchar 'NULL'.
Following syntax works perfectly:
INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('INSERT INTO proposition VALUES(',NEW.id,',',quote_nullable(NEW.text),',',CASE WHEN quote_nullable(NEW."categorySignId") IS NULL THEN 'NULL' ELSE NEW."categorySignId" END,');'))

Is there any risk with an insert query which would looks like: INSERT INTO proposition VALUES(1,'Exemple','1');
While third parameter is an integer?

#2 Re : Général » audit table with integer column » 22/07/2016 17:05:36

I already tried this syntax but postgres doesn't like:
mydb=# INSERT INTO proposition VALUES(2944,'Example',NULL);
ERROR:  invalid input syntax for integer: "NULL"
LINE 1: ...),',',CASE WHEN NEW."categorySignId" IS NULL THEN 'NULL' ELS...
                                                             ^
QUERY:  INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('INSERT INTO proposition VALUES(',NEW.id,',',quote_nullable(NEW.text),',',CASE WHEN NEW."categorySignId" IS NULL THEN 'NULL' ELSE NEW."categorySignId" END,');'))
CONTEXT:  PL/pgSQL function process_audit() line 269 at SQL statement

This comes from the apostrophes around 'NULL'

#3 Général » audit table with integer column » 22/07/2016 10:28:15

massup
Réponses : 4

Hell,

I have a table proposition
     Column       |  Type     |                        Modifiers
------------------+---------+-----------------------------------------------------
id                    | integer  | not null default nextval('proposition_id_seq'::regcl
text                 | text      |
categorySignId | integer  |

I'm auditing in table audit all modifications made on table proposition via an audit function containing:
      ELSIF (TG_TABLE_NAME = 'proposition') THEN
        IF (TG_OP = 'DELETE') THEN
            INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('DELETE FROM proposition WHERE id=',OLD.id,';'));
            RETURN OLD;
        ELSIF (TG_OP = 'UPDATE') THEN
            INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('UPDATE proposition SET text=',quote_nullable(NEW.text),',"categorySignId"=',NEW."categorySignId",' WHERE id=',NEW.id,';'));
            RETURN NEW;
        ELSIF (TG_OP = 'INSERT') THEN
            INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('INSERT INTO proposition VALUES(',NEW.id,',',quote_nullable(NEW.text),',',NEW."categorySignId",');'));
            RETURN NEW;
        END IF;

When i insert following record in proposition:
INSERT INTO proposition VALUES(1,'Exemple',NULL);

I obtain following an audit record with following query field:
INSERT INTO proposition VALUES(1,'Exemple',);

This query is not correct, is there a way to obtain NULL string when the integer column is NULL in order to rebuild initial query?

#4 Re : PL/pgSQL » Automic replacement in SQL query » 21/07/2016 09:21:43

Apologize for this.
I might have missed an update when doing my test, you are right this is working fine.
Thanks for your help.

#5 Re : PL/pgSQL » Automic replacement in SQL query » 20/07/2016 15:05:22

Let's consider:
-table test containing fields id and description.
-Table audit containing fields id, timestamp, query
-The audit_function:

CREATE OR REPLACE FUNCTION audit_function() RETURNS TRIGGER AS $audit$
    BEGIN
        --
        -- Create a row audit to reflect the operation performed on test,
        --
          IF (TG_TABLE_NAME = 'test') THEN
        IF (TG_OP = 'DELETE') THEN
            INSERT INTO audit (stamp,query) VALUES(now(), CONCAT('DELETE FROM test WHERE id=',OLD.id,';'));
            RETURN OLD;
        ELSIF (TG_OP = 'UPDATE') THEN
            INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('UPDATE test SET id=', NEW.id, ',description=', quote_literal(NEW.description)';'));
            RETURN NEW;
        ELSIF (TG_OP = 'INSERT') THEN
            INSERT INTO audit (stamp,query) VALUES( now(), CONCAT('INSERT INTO test VALUES(',NEW.id,',',quote_literal(NEW.description),');'));
            RETURN NEW;
        END IF;
      RETURN NULL;
    END;
$audit$ LANGUAGE plpgsql;

-Trigger
CREATE TRIGGER test_audit
AFTER INSERT OR UPDATE OR DELETE ON test
    FOR EACH ROW EXECUTE PROCEDURE audit_function();

If i insert record id=1, description=Here's the example (INSERT INTO test (id,description) VALUES (1,'Here''s the example');)
I will get in audit table a record with query=INSERT INTO test (stamp,query) VALUES (yyyymmdd-hhmmss,'Here's the example');

#6 Re : PL/pgSQL » Automic replacement in SQL query » 20/07/2016 09:21:01

It works in your example.
But how can i manage it in the case of a variable new.text containing << doesn't  >> with no backslash?
quote_nullable(new.text) doesn't return expected result.

#7 Re : PL/pgSQL » Automic replacement in SQL query » 20/07/2016 08:25:25

Hello,

I tried with quote_nullable and quote_literal but it doesn't produces expected result.
For example i have a variable new.text containing txt : doesn't.
quote_nullable(new.text) returns 'doesn't' while i would have expected to get 'doesn''t'
I obtain same result with quote_literal.

#8 PL/pgSQL » Automic replacement in SQL query » 18/07/2016 16:22:39

massup
Réponses : 9

Hello,
In order to audit all changes made on a database and reapply them when necessary on an other copy of this db, i have created triggers on related tables, an audit table and a function which writes in audit table each time a change occures on other tables.
I'm reconstructing the SQL query when writting it in audit.
The problem is with text strings containing apastrophe. I would like to automaticaly replace single apastrophe by double apastrophe in order to have a correct syntax.
Below example of the syntax i'm using:

CREATE OR REPLACE FUNCTION process_audit() RETURNS TRIGGER AS $audit$
    BEGIN
        --
        -- Create a row in auditv2 to reflect the operation performed on tablename,
        -- make use of the special variable TG_OP to work out the operation.
        --
          IF (TG_TABLE_NAME = 'tablename') THEN
        IF (TG_OP = 'DELETE') THEN
            INSERT INTO auditv2 (stamp,query) VALUES(now(), CONCAT('DELETE FROM tablename WHERE id=',OLD.id,';'));
            RETURN OLD;
        ELSIF (TG_OP = 'UPDATE') THEN
            INSERT INTO auditv2 (stamp,query) VALUES( now(), CONCAT('UPDATE tablename SET "typeId"=', NEW."typeId", ',subgroup=''', NEW.subgroup, ''',name=''', NEW.name, ''',"imageName"=''', NEW."imageName", ''' WHERE id=', NEW.id,';'));
            RETURN NEW;
        ELSIF (TG_OP = 'INSERT') THEN
            INSERT INTO auditv2 (stamp,query) VALUES( now(), CONCAT('INSERT INTO tablename VALUES(',NEW.id,',''', NEW."typeId",''',''',NEW.subgroup,''',''',NEW.name,''',''',NEW."imageName",''');'));
            RETURN NEW;

If name field contains for example "It's the example" i would like to replace it by "It''s the example".

I tried with REPLACE command ( REPLACE (NEW.name, "'", "''") but when inserting a record SQL tells me column "'" doesn't exist!

Any idea?

Pied de page des forums

Propulsé par FluxBB