it's possible to define an alias for a stored procedure (sql-native)?
In the documentation, reference is made only to: SEQUENCE and TABLES
1) CREATE PROCEDURE OWXXCOLL.STORED1()
2) CREATE ALIAS DB2C.STORED1 FOR OWXXCOLL.STORED1;
3) CALL DB2C.STORED1();
EDIT 2021-05-14
The original question arises for the following problem (I was hoping to get away with using aliases)
Intro
I have defined a native SP
The OWXXCOLL schema is the same one I also use for tables/index...
(I noticed that the tables also have different aliases)
CREATE PROCEDURE OWXXCOLL.STORED1(...)
LANGUAGE SQL
ISOLATION LEVEL CS
WLM ENVIRONMENT FOR DEBUG MODE WLMENV1
ALLOW DEBUG MODE
BEGIN
...
END#
I also modified the cobol program (name:PGMSTO1) to call the Stored with the CALL statement (without qualifier)
EXEC SQL
CALL STORED1 (...)
END-EXEC.
The problem
The various table accesses (SELECT) work correctly BUT When I run the PGMSTO1 the call to the Stored ends with sqlcode -440
NO AUTHORIZED PROCEDURE BY THE NAME STORED1 HAVING COMPATIBLE ARGUMENTS WAS FOUND<
The error comes from the fact that it is not using the owner OWXXCOLL but DB2C (DB2C is user who scheduled the jcl/batch)
If I enter the qualifier (OWXXCOLL) the call it's OK.
I don't understand what to check and what configurations are missing.
Thanks
The cobol program (PGMSTO1) that calls the stored procedure has the following BIND parameters:
COLLID NAME OWNER CREATOR QUALIFIER DYNAMICRULES PATHSCHEMAS
OWXXCOLL PGMSTO1 DB2C DB2C FPXX B "DB2C"
"DB2C" path is used to resolve unqualified stored procedure
I will need to modify the bind parameter "PATH".
I will try to ask to add my schema as well (OWXXCOLL)
PATH("OWXXCOLL","DB2C")
Related
I am working on a stored procedure to update values in a couple of tables in one go. I am getting CLI0118E Invalid SQL syntax. during the execution. However, I am not able to find the issue with my stored procedure as I am not quite familiar with the whole stored procedure rules. Any help on this would be much appreciated.
CREATE OR REPLACE PROCEDURE MYPROCNAME(IN VAL1 VARCHAR(255),IN VAL2 VARCHAR(255),IN VAL3 VARCHAR(255),IN VAL4 VARCHAR(255))
LANGUAGE SQL
BEGIN ATOMIC
UPDATE SCHEMA1.TABLE1 SET col1=VAL2, col2=VAL3, col3=VAL4 WHERE col1=VAL1;
UPDATE SCHEMA2.TABLE1 SET col1=VAL3, col2=VAL3, col3=VAL4 WHERE col1=VAL1;
FOR v AS
SELECT col1 FROM SCHEMA2.TABLE2 WHERE col2=VAL1
DO
UPDATE SCHEMA2.TABLE3 SET col2=VAL2 WHERE col1=v.col1 AND col3='<<value>>';
UPDATE SCHEMA2.TABLE3 SET col2=VAL3 WHERE col1=v.col1 AND col3='<<value2>>';
END FOR;
END
#
In the FOR statement, I even tried like this FOR v AS cur1 CURSOR FOR
Used this to compile the procedure first -- db2 -td# -vf myprocname.db2
Then, tried executing the procedure using -- db2 call MYPROCNAME 'val1','val2','val3','val4'
Also, tried the execution in diff combinations like db2 call MYPROCNAME ('val1','val2','val3',val4'), db2 call MYPROCNAME("val1","val2","val3","val4"), etc...
If I use () in the db2 call, I get the bash: syntax error near unexpected token (' error.
PS: I am using db2 11.5
As mentioned in the comment to my question above, running the following db2 "call MYPROCNAME ('val1', 'val2', 'val3', 'val4')" fixes the issue
We use APIs, baninst1.PP_DEDUCTION.p_update and baninst1.PP_DEDUCTION.p_create, to maintain our payroll tables of benefit/deduction data. Numerous packages utilize the APIs. We would like to create a package containing the API call that all the existing packages can use and remove the code that is repeated in each package. I tried EXECUTE IMMEDIATE for the purpose of having a dynamic API name. However, I have not been able to get the syntax correct. I’m hoping you will help me.
create or replace PACKAGE BODY "ORBIT"."MM_BENEFITS_COMMON" AS
PROCEDURE PAY_P_EMPLOYEE_BENEFIT_ACTION(pi_benefit_action IN VARCHAR2,
pi_pidm IN pdrbded.pdrbded_pidm%TYPE,
pi_status IN pdrdedn.pdrdedn_status%TYPE,
pi_bdca_code IN pdrbded.pdrbded_bdca_code%TYPE,
pi_effective_date IN pdrdedn.pdrdedn_effective_date%TYPE DEFAULT NULL, pi_user_id IN pdrdedn.pdrdedn_user_id%TYPE DEFAULT NULL, pi_data_origin IN pdrdedn.pdrdedn_data_origin%TYPE DEFAULT NULL,
po_base_rowid_out OUT gb_common.internal_record_id_type,
po_detail_rowid_out OUT gb_common.internal_record_id_type,
pi_amount1 IN pdrdedn.pdrdedn_amount1%TYPE DEFAULT NULL,
pi_opt_code1 IN pdrdedn.pdrdedn_opt_code1%TYPE DEFAULT NULL) IS
BEGIN
--Call the API for p_create or p_update.
baninst1.PP_DEDUCTION.pi_benefit_action(p_pidm => pi_pidm,
p_status => pi_status,
p_bdca_code => pi_bdca_code,
p_effective_date => CASE
WHEN pi_benefit_action 'p_create' THEN
TRUNC(pi_begin_date)
ELSE
TRUNC(pi_effective_date)
END,
p_user_id => pi_user_id,
p_data_origin => pi_data_origin,
p_base_rowid_out => po_base_rowid_out,
p_detail_rowid_out => po_detail_rowid_out, p_amount1 => pi_amount1,
p_opt_code1 => CASE
WHEN LENGTH(pi_opt_code1) = 1 THEN
'0' || pi_opt_code1
ELSE pi_opt_code1
END);
END PAY_P_EMPLOYEE_BENEFIT_ACTION;
END MM_BENEFITS_COMMON;
create or replace PACKAGE BODY "ORBIT"."MM_BENEFIT_TEST" AS
PROCEDURE PAY_P_MM_BENEFIT_TEST IS
lv_base_rowid_out gb_common.internal_record_id_type;
lv_detail_rowid_out gb_common.internal_record_id_type;
BEGIN
--Pass data to the common benefits package for the api call
MM_BENEFITS_COMMON.PAY_P_EMPLOYEE_BENEFIT_ACTION('p_update', 9999999, 'A', 'VI1', '01-JAN-2022', 'MM_Test', 'MM_TEST', lv_base_rowid_out, lv_detail_rowid_out, 25.82, NULL);
END PAY_P_MM_BENEFIT_TEST;
END MM_BENEFIT_TEST;
I'm not sure what's bothering you, actually. You did post some code, but - I don't know what it represents.
Let's see what you said:
"We use APIs, baninst1.PP_DEDUCTION.p_update and baninst1.PP_DEDUCTION.p_create, to maintain our payroll tables of benefit/deduction data."
OK
"Numerous packages utilize the APIs."
it means that there are many packages and they call those p_update and p_create procedures; that's also OK
"We would like to create a package containing the API call that all the existing packages can use and remove the code that is repeated in each package."
that would be a new package; you'd cut that piece of code from all of your packages and paste it into a new one.
OK, makes sense. Instead of all that code (in every package), you'd put call to newly created procedures (that reside in a newly created package)
"I tried EXECUTE IMMEDIATE for the purpose of having a dynamic API name. However, I have not been able to get the syntax correct."
why dynamic SQL? There's nothing dynamic here. Instead of dozens of lines of code you currently have (that do something), you'd put one line - the one that calls that newly created procedure (and pass parameters)
From my point of view, there's nothing unusual in what you want to do and I can't imagine what problems you could have in doing it; it's pretty much straightforward.
CREATE OR REPLACE PROCEDURE wh.sp_schema.my_sp(arg1 STRING, arg2 STRING)
RETURNS VARCHAR(16777216)
LANGUAGE JAVASCRIPT
AS
$$
stmt = snowflake.createStatement(
{sqlText: `CREATE OR REPLACE TABLE wh.table_schema.RAW_`+arg2+`_`+arg1+` LIKE wh.temp_schema.RAW_`+arg2+`_TEMPLATE`}
);
rs = stmt.execute();
rs.next();
output = rs.getColumnValue(1);
return output;
$$
;
when i create the above stored procedure - below message is displayed
Function my_sp successfully created.
when i run the stored procedure from the worksheet using
CALL my_sp('2018','abc');
I see the below error:
SQL compilation error: Invalid identifier my_sp
I even tried running using fully qualified name for the stored procedure:
CALL wh.sp_schema.my_sp('2018','abc');
I still see the below error:
SQL compilation error: Invalid identifier my_sp.
Also, i would like to know a command to see if my stored procedure has been created or not.
I've looked in the snowflake UI, under 'wh" warehouse and "sp_schema" schema, but could not see anything(stored procedures) other than tables.
This is a simple case of a grants issue. As SECURITYADMIN, if you apply the necessary grants to the role you're using,
GRANT USAGE ON DATABASE wh TO ROLE <role>;
GRANT USAGE ON SCHEMA wh.sp_schema TO ROLE <role>;
GRANT USAGE ON ALL PROCEDURES IN SCHEMA wh.sp_schema TO ROLE <role>;
you'll get past this ambiquous error and get to the next real error
JavaScript execution error: Uncaught ReferenceError: arg2 is not defined in MY_SP
Which is due to javascript SP parameters being case sensitive, so you need to modify the code as follows (change arg to ARG):
{sqlText: `CREATE OR REPLACE TABLE wh.table_schema.RAW_`+ARG2+`_`+ARG1+` LIKE wh.temp_schema.RAW_`+ARG2+`_TEMPLATE`}
Snowflake does not support Stored Procedures. They are implementingjavascript` based APIs for it, but it's still in draft mode.
Here's the link to their APIs (Note Draft Document is written on right hand side).
and here's the link to a ticket on their support forum where they've confirmed it.
There are two obvious possibilities why this may go wrong:
You've implicitly donated the procedure ownership to a ROLE you don't have
There is now a mismatch between the procedure and call type signatures. Verify that there is a wh.sp_schema.my_sp(STRING, STRING) and that your call is like CALL wh.sp_schema.my_sp('2018'::STRING,'abc'::STRING);
Try to solve the error by following query syntax.
CALL database_name.schema.MY_SP('database_name','schema');
How would you run a stored procedure via DB2 command of QShell, as I need a simple way to unit test a change to a stored procedure?
On IBM i
Started qshell with
QSH
Entered these db2 commands in various formats without success
db2 call libraryname.stroredprocedurename('param1value' 'param2value' ?)
db2 call libraryname.stroredprocedurename ('param1value' 'param2value' ?)
db2 call libraryname.stroredprocedurename (param1value param2value ?)
Only reference source I could find
Have tried using CLP to call stored proceduresbut there are no CLP examples
here
You can also use the JDBC client included with jt400.jar. You can run it from QSH by using the following command.
java -cp /qibm/proddata/os400/jt400/lib/jt400.jar com.ibm.as400.access.jdbcClient.Main jdbc:db2:localhost
The client will also handle stored procedure output parameters as shown by the following example.
create procedure add1(in inparm int, out outparm int) language sql begin set outparm = inparm +1; end
call add1(1,?)
Parameter 1 returned 2
I use Squirrel SQL Client http://squirrel-sql.sourceforge.net/ to test all of my SQL.
call libraryname.stroredprocedurename('param1value', 'param2value')
Note the separator is a comma ,, and ? is not a valid parameter marker when called interactively like this.
In addition to the client, you will need a JDBC driver. You can use the JTOpen driver for IBM i found here: http://jt400.sourceforge.net/
From the CL command line, or inside a CL program, you can use the RUNSQL command to execute an SQL statement.
If you're going to be trying multiple ad hoc SQL statements, you might use the STRSQL command. Personally, I tend to use the SQL window provided as part iNavigator.
Finally got the basic syntax
db2 "CALL lib.proc ('parmvalue1')"
Which resulted in:
DB20000I THE SQL COMMAND COMPLETED SUCCESSFULLY.
I am working on an application built in Delphi 2010 that uses UIB to connect to a Firebird 2.5 database. The application has been running using the default character set for a long time, i.e. nobody gave character sets any special thought and it has simply been working. Currently I am trying to make it work correctly with UTF-8 data.
In doing so I have hit upon a problem with TUIBQuery and parameterized queries. When using Database.Charset=csUTF8 and setting a parameter value for a CHAR(n)-field and retrieving it before executing the query the value is truncated.
Unfortunately some of my code writes and reads parameter like this in a number of places and therefore dies an ugly death.
To isolate and demonstrate the problem I created a simple fresh database with DEFAULT CHARACTER SET UTF-8 and a table like this:
CREATE TABLE TEST (
CHARFIELD CHAR(20),
VARCHARFIELD VARCHAR(20)
);
I set up the application to connect to the database using TUIBDatabase and TUIBTransaction. Then I created a TUIBQuery-instance, set SQL to a parameterized INSERT-statement into this table, and set the parameters:
Query := TUIBQuery.Create(NIL);
Query.Transaction := Transaction;
Query.SQL.Text := 'INSERT INTO TEST (CHARFIELD, VARCHARFIELD) VALUES (:CHARFIELD, :VARCHARFIELD)';
Query.Prepare(True);
s:= 'ABC';
Query.Params.ByNameAsString['CHARFIELD'] := s;
Query.Params.ByNameAsString['VARCHARFIELD'] := s;
When I now read the parameter-values back like this:
s := Query.Params.ByNameAsString['CHARFIELD'];
s := Query.Params.ByNameAsString['VARCHARFIELD'];
The results are correct for Database.Charset=csNone. But when I instead specify DataBase.Charset=csUTF8 the value for CHARFIELD is truncated to 'A' instead of 'ABC'. The value for VARCHARFIELD is fine. The behaviour is independent of the actual data, I do not have to actually use non-ASCII-characters to provoke it, as the sample shows.
Calling ExecSQL() on the query works correctly and INSERTs the data as expected in both cases.
I have uploaded sourcecode to my simple test program as UIB_UTF8_Test.zip.
Does someone here have any idea what I may be doing wrong and how to do it right?