I am a newbie in the area of redshift data modeling and got myself into trouble with an error.ERROR:
--Final version
syntax error ERROR: operator does not exist: text | record Hint: No
operator matches the given name and argument type(s). You may need to
add explicit type casts. Where: SQL statement "SELECT 'create temp
table ' || $1 || ' as select * from' | $2 |" PL/pgSQL function "egen"
line 36 at execute statement [ErrorId:
1-61dc32bf-0a451f5e2c2639235abb8876]
I am trying to do a simple transformation that gets returned in output when the procedure is called. (As of now I got to find from the documentation we have to use either temp table or cursors to achieve this)
Pseudocode:
I am trying to restrict data to its latest one in (2019) Get the
list of managers create columns if a person is a manager or not from the list.
Return it as a result
Data looks as follows Employee Data
My Select query works fine out of the procedure, please find my complete code below.
CREATE OR REPLACE PROCEDURE EGEN(tmp_name INOUT varchar(256) )
AS $$
DECLARE
--As i have less data managed to create it as an array or please use temp or table and join it with the actual query to perform transformation
MGR_RECORD RECORD;
DATAS RECORD;
item_cnt int := 0;
V_DATE_YEAR int := 0;
BEGIN
--EXECUTE (select cast(extract(year from current_date) as integer)-3) INTO V_DATE_YEAR;
--Manager Records are stored here below
SELECT DISTINCT managerid from "dev"."public"."emp_salary" INTO MGR_RECORD;
SELECT employeeid,
managerid,
promotion,
q_bonus,
d_salary,
case when contractor = 'x'
then 'TemporaryEmployee'
else 'PermanentEmployee'
END as EmployeeType,
-- IFstatement not supported under select query
case when employeeid in (select distinct managerid FROM "dev"."public"."emp_salary" )
then 'Manager'
else 'Ordinary FTE'
END as FTETYPE
FROM "dev"."public"."emp_salary" where cast(extract(year from promotion) as int ) >= 2019 into DATAS;
--COMMIT;
tmp_name := 'ManagerUpdatedTable';
EXECUTE 'drop table if exists ' || tmp_name;
EXECUTE 'create temp table ' || 'ManagerUpdatedTable' || ' as select * from' |DATAS| ;
END;
$$ LANGUAGE plpgsql;
-- Call tests CALL EGEN('myresult'); SELECT * from myresult;
Also, additional query (Can we replace )
case when employeeid in (select distinct managerid FROM "dev"."public"."emp_salary" )
then 'Manager'
else 'Ordinary FTE'
END as FTETYPE
this transform in query to IF , if possible please provide details.
Thanks and Regards,
Gabby
Related
I need to find out rows that are present in table A and missing from table B (using LEFT JOIN) wherein table A and table B are two tables with same structure but within different schema.
But the query has to be constructed using Dynamic SQL and the columns that need to be used for performing JOIN are stored in a string. How to extract the column names from string and use them to dynamically construct below query :
Database is Azure SQL Server
eg :
DECLARE #ColNames NVARCHAR(150) = 'col1,col2'
Query to be constructed based on columns defined in ColNames :-
SELECT *
FROM Table A
Left Join
Table B
ON A.col1 = B.col1
AND A.col2 = B.col2
AND B.col1 IS NULL AND B.col2 IS NULL
If the number of columns in #ColNames is more then the SELECT statement needs to cater for all the column.
Without knowing the full context, try this:
DECLARE #ColNames NVARCHAR(150) = 'col1,col2'
DECLARE #JoinContion NVARCHAR(MAX) = ''
DECLARE #WhereCondition NVARCHAR(MAX) = ''
SELECT #JoinContion += CONCAT('[a].', QUOTENAME(Value), ' = ', '[b].', QUOTENAME(Value), (CASE WHEN LEAD(Value) OVER(ORDER BY Value) IS NOT NULL THEN ' AND ' ELSE '' END))
,#WhereCondition += CONCAT('[a].', QUOTENAME(Value), ' IS NULL', (CASE WHEN LEAD(Value) OVER(ORDER BY Value) IS NOT NULL THEN ' AND ' ELSE '' END))
FROM STRING_SPLIT(#ColNames,N',')
SELECT #JoinContion, #WhereCondition
String_Split: To split the input string into columns
Lead: to determine if we need the AND keyword when it's not the last row.
Be aware the NOT EXISTS is probably a better solution then LEFT JOIN
I’m currently trying to implement Postgres text search for my rails app. My migrations run successfully creating all the tables. One of my migrations adds a trigger to update a ts_vector column (TSV) on my listings table. It does a bunch of joins on tables and two joins on intermediate tables. Not sure if this trigger is the most efficient but it works when it’s called by a procedure.
CREATE OR REPLACE FUNCTION update_listings_tsv() RETURNS trigger AS $$
BEGIN
NEW.tsv := (
SELECT
setweight(to_tsvector(l.item_name), 'A') ||
setweight(to_tsvector(l.description), 'B') ||
setweight(to_tsvector(categories.name), 'B') ||
setweight(to_tsvector(sub_categories.name), 'B') ||
setweight(to_tsvector(sizes.name), 'B') ||
setweight(to_tsvector(users.username), 'C') ||
setweight(to_tsvector(string_agg(DISTINCT brands.name, ',')), 'A') ||
setweight(to_tsvector(string_agg(DISTINCT colours.name, ',')), 'B')
FROM listings l
JOIN users ON users.id = l.user_id
JOIN categories ON categories.id = l.category_id
JOIN sub_categories ON sub_categories.id = l.sub_category_id
JOIN sizes ON sizes.id = l.size_id
JOIN conditions ON conditions.id = l.condition_id
JOIN brands_listings ON brands_listings.listing_id = l.id
JOIN brands ON brands.id = brands_listings.brand_id
JOIN colours_listings ON colours_listings.listing_id = l.id
JOIN colours ON colours.id = colours_listings.colour_id
WHERE l.id = NEW.id
GROUP BY l.id, users.id, categories.id, sub_categories.id, sizes.id, conditions.id
);
RETURN NEW;
END;
$$ LANGUAGE plpgsql;
This is where the problem is:
The update procedure that calls the trigger works fine. As intended when a listing column updates the TSV column for that exact listing also updates.
CREATE TRIGGER tsvectorupdateupdate BEFORE UPDATE
ON listings FOR EACH ROW EXECUTE PROCEDURE update_listings_tsv();
The insert procedure doesn’t work. I’ve tried it with a BEFORE and AFTER INSERT. I don’t know where the problem is. After a db:reset the TSV column should already be populated with a ts_vector value.
CREATE TRIGGER tsvectorupdateinsert AFTER INSERT OR UPDATE
ON listings FOR EACH ROW EXECUTE PROCEDURE update_listings_tsv(tsv);
If you do an AFTER trigger, it is too late to change the value. If you do a BEFORE trigger, it is too early to join against the (not yet existing) row in the table "listings". You need to get the new data from NEW, not from "listings".
I've reduced this to its minimal essence, fleshing it back out with extra tables and columns is up to you:
CREATE OR REPLACE FUNCTION update_listings_tsv() RETURNS trigger AS $$
BEGIN
NEW.tsv := (
SELECT
setweight(to_tsvector(NEW.item_name), 'A') ||
setweight(to_tsvector(users.username), 'C')
FROM users where users.id = NEW.user_id
);
RETURN NEW;
END;
$$ LANGUAGE plpgsql;
CREATE TRIGGER tsvectorupdateinsert BEFORE INSERT OR UPDATE
ON listings FOR EACH ROW EXECUTE PROCEDURE update_listings_tsv();
I try to check existence of a record before inserting into table.
SET TERM ^ ;
CREATE PROCEDURE add_videorecord(tab_name varchar(31), col_name varchar(31),
col_value varchar(100))
RETURNS (status int)
AS
BEGIN
status=1;
if (not exists(
select * from :tab_name where :col_name = :col_value))
then
execute statement 'insert into "'||:tab_name||'" ("'||:col_name||'") values("'||:col_value||'")';
else
status=0;
END^
SET TERM ; ^
And get FlameRobin error:
Message: isc_dsql_prepare failed
SQL Message : -104
Invalid token
Engine Code : 335544569
Engine Message :
Dynamic SQL Error
SQL error code = -104
Token unknown - line 10, column 15
:
Why it's unknown token? I try to use the input parameter.
You can't directly parametrize an object name (like a table name, column name, etc) like you do in your select. Parameters can only be used for values.
If you want to parametrize an object name, you need to concatenate it into a query string, like you already do for your insert statement. You just need to watch out for SQL injection (eg by checking the name against a known set of accepted object names).
As an example (without checking for valid table and column name!):
CREATE PROCEDURE ADD_RECORD (
TAB_NAME VARCHAR(31),
COL_NAME VARCHAR(31),
COL_VALUE VARCHAR(100) )
RETURNS (
STATUS INTEGER )
AS
DECLARE temp INTEGER;
BEGIN
status=1;
execute statement
('select 1 from "' || tab_name || '" where "' || col_name || '" = ?') (:col_value)
into :temp;
if (temp is null)
then
execute statement
('insert into "'|| tab_name||'" ("' || col_name|| '") values(?)') (:col_value);
else
status=0;
END
Using a single (dynamically created) MERGE statement, you might be able to simplify this further.
I have a need to run a recursive CTE within a stored proc, but I can't get it past this:
SQL0104N An unexpected token "with" was found following "SET count=count+1;
". Expected tokens may include: "". LINE NUMBER=26.
My google-fu showed a couple of similar topics, but none with resolution.
The query functions as expected outside of the stored proc, so I'm hoping that there's some syntactic sugar I'm missing that'll let this work. Similarly, the proc compiles and works without the query.
Here's a contrived example:
--setup
create table tree (id integer, name varchar(50), parent_id integer);
insert into tree values (1, 'Alice', null);
insert into tree values (2, 'Bob', 1);
insert into tree values (3, 'Charlie', 2);
-
- the proc
create or replace procedure testme() RESULT SETS 1 LANGUAGE SQL
BEGIN
DECLARE SQLSTATE CHAR(5);
DECLARE SQLCODE integer default 0;
DECLARE count INTEGER;
DECLARE sum INTEGER;
DECLARE total INTEGER;
DECLARE id INTEGER;
DECLARE curs CURSOR WITH RETURN FOR
select count,sum from sysibm.sysdummy1;
DECLARE hiercurs CURSOR FOR
select id from tree order by id;
SET bomQuery='';
PREPARE stmt FROM bomQuery;
SET count = 0;
SET sum = 0;
set total = 0;
OPEN hiercurs;
FETCH hiercurs INTO id;
WHILE (SQLCODE <> 100) DO
SET count=count+1;
with org (level,id,name,parent_id) as
(select 1 as level,root.id,root.name,root.parent_id from tree root where root.id=id
union all
select level+1,employee.id,employee.name,employee.parent_ id from org boss, tree employee
where level < 5 and employee.parent_id=boss.id)
select count(1) into sum from org;
SET total=total+sum;
FETCH hiercurs INTO id;
END WHILE;
CLOSE hiercurs;
OPEN curs;
END
the cte in db2 doesn't seem to recognize the scalar result of the query, and so it won't let the select into work (not a problem on Oracle or SQLServer)...solution is to open a cursor and FETCH INTO (instead of SELECT INTO) instead.
In addition to rjb's suggestion of enclosing the CTE query inside a cursor, you can also stuff the CTE into a user-defined function or a view, and then code a straight select against that object into your stored procedure.
i'm creating a procedure to update/insert a table using merge statement(upsert).now i have a problem: using procedure parameters i have to do this upsert.
procedure xyz( a in table.a%type,b in table.b%type,....)
is
some local variables;
begin
merge into target_table
using source_table --instead of the source table, i have to use procedure parameters here
on (condition on primary key in the table)
when matched then
update the table
when not matched then
insert the table ;
end xyz;
so how to use procedure parameters instead of source table in merge statement?? or
suggest me a query to fetch the procedure parameters and use it as source table values.
help me please.
Thanks in advance.
I know that I'm eight years late to the party, but I think that I was trying to do something similar to what you were doing, but trying to Upsert based on parameters passed into a stored procedure that returns an empty string on success and an error on failure back to my VB Code. Below is all of my code along with comments explaining what I did, and why I did it. Let me know if this helps you or anyone else. This is my first time answering a post.
PROCEDURE UpsertTSJobData(ActivitySeq_in IN NUMBER,
Owner_in In VARCHAR2,
NumTrailers_in IN NUMBER,
ReleaseFormReceived_in IN NUMBER,
Response_out OUT VARCHAR2) AS
err_num NUMBER;
err_msg VARCHAR2(4000);
BEGIN
--This top line essentially does a "SELECT *" from the named table
--and looks for a match based on the "ON" statement below
MERGE INTO glob1app.GFS_TS_JOBDATA_TAB tsj
--This select statement is used for the INSERT when no match
--is found and the UPDATE when a match is found.
--It creates a "pseudo-table"
USING (
SELECT ActivitySeq_in AS ActSeq,
Owner_in As Owner,
NumTrailers_in As NumTrailers,
ReleaseFormReceived_in As ReleaseFormReceived
FROM DUAL) input
--This ON statement is what we're doing the match on to find
--matching records. This decides whether it will be an
--INSERT or an UPDATE
ON (tsj.Activity_seq = ActivitySeq_in)
WHEN MATCHED THEN
--Here we UPDATE based on the passed in input table
UPDATE
SET OWNER = input.owner,
NUMTRAILERS = input.NumTrailers,
RELEASEFORMRECEIVED = input.releaseformreceived
WHEN NOT MATCHED THEN
--Here we INSERT based on the passed in input table
INSERT (
ACTIVITY_SEQ,
OWNER,
NUMTRAILERS,
RELEASEFORMRECEIVED
)
VALUES (
input.actseq,
input.owner,
input.numtrailers,
input.releaseformreceived
);
Response_out := '';
EXCEPTION
WHEN OTHERS THEN
err_num := SQLCODE;
err_msg := SUBSTR(SQLERRM, 1, 3900);
Response_out := TO_CHAR (err_num) || ': ' || err_msg;
END;
Maby something like
DECLARE V_EXISTS NUMBER;
BEGIN SELECT COUNT(*) INTO V_EXISTS FROM TARGET_TABLE WHERE PK_ID = :ID;
IF V_EXISTS > 0 THEN
-- UPDATE
ELSE
-- INSERT
END IF;
END;
Also, you may try to use so-called tempotary table (select from DUAL)
CREATE TABLE TEST (N NUMBER(2), NAME VARCHAR2(20), ADRESS VARCHAR2(100));
INSERT INTO TEST VALUES(1, 'Name1', 'Adress1');
INSERT INTO TEST VALUES(2, 'Name2', 'Adress2');
INSERT INTO TEST VALUES(3, 'Name3', 'Adress3');
SELECT * FROM TEST;
-- test update
MERGE INTO TEST trg
USING (SELECT 1 AS N, 'NameUpdated' AS NAME,
'AdressUpdated' AS ADRESS FROM Dual ) src
ON ( src.N = trg.N )
WHEN MATCHED THEN
UPDATE
SET trg.NAME = src.NAME,
trg.ADRESS = src.ADRESS
WHEN NOT MATCHED THEN
INSERT VALUES (src.N, src.NAME, src.ADRESS);
SELECT * FROM TEST;
-- test insert
MERGE INTO TEST trg
USING (SELECT 34 AS N, 'NameInserted' AS NAME,
'AdressInserted' AS ADRESS FROM Dual ) src
ON ( src.N = trg.N )
WHEN MATCHED THEN
UPDATE
SET trg.NAME = src.NAME,
trg.ADRESS = src.ADRESS
WHEN NOT MATCHED THEN
INSERT VALUES (src.N, src.NAME, src.ADRESS);
SELECT * FROM TEST;
DROP TABLE TEST;
see here
Its very difficult to tell from you question exactly what you what, but I gather you want the table that you are merging into ( or on ) to be dynamic. In that case, what you should be using is the DBMS_SQL package to create dynamic SQL