I am trying to create a function on Postgres 10 that would check whether all the given ids exist
in the target table.
CREATE OR REPLACE FUNCTION "public"."has_mismatching_ids"(
_ids TEXT[],
_column_name TEXT,
_table_name TEXT,
_schema_name TEXT DEFAULT 'public'
)
RETURNS BOOLEAN AS
$$
DECLARE
result TEXT[] := '{}';
_query TEXT := '';
BEGIN
IF _ids IS NULL THEN
RETURN false;
END IF;
_query := CONCAT('SELECT UNNEST($1) EXCEPT SELECT "', _column_name, '" FROM "', _schema_name ,'"."', _table_name ,'"');
EXECUTE _query INTO result USING $1;
RAISE NOTICE 'Executed';
IF array_length(result, 1) > 0 THEN
RETURN true;
END IF;
RETURN false;
END;
$$
LANGUAGE plpgsql
STABLE
SECURITY INVOKER;
This is my function that returns correctly true if all the ids exist, but it throws an exception if any of
the ids does not exist.
CREATE TABLE "public"."categories" ("category_id" TEXT NOT NULL PRIMARY KEY);
INSERT INTO "public"."categories" ("category_id") VALUES ('foo');
INSERT INTO "public"."categories" ("category_id") VALUES ('bar');
This works
SELECT FROM "public"."has_mismatching_ids"(ARRAY['foo'], 'category_id', 'categories', 'public');
NOTICE: Executed
has_mismatching_ids
---------------------
f
(1 row)
but this fails to even
SELECT FROM "public"."has_mismatching_ids"(ARRAY['foo2'], 'category_id', 'categories', 'public');
ERROR: malformed array literal: "foo2"
DETAIL: Array value must start with "{" or dimension information.
CONTEXT: PL/pgSQL function public.has_mismatching_ids(text[],text,text,text) line 14 at EXECUTE
and strangely this works also if I write a function that uses explicit names and without EXECUTE
SELECT UNNEST(ARRAY['foo', 'foo3']) EXCEPT SELECT "category_id" FROM "public"."categories";
unnest
--------
foo3
(1 row)
Why doesn't my dynamic function work when comparing to values that do not exist in the target table?
You may use a simple SQL to compare if all array element exists.
CREATE OR replace FUNCTION "public"."has_mismatching_ids"(
_ids text[],
_column_name text,
_table_name text,
_schema_name text DEFAULT 'public' )
returns boolean AS $$
DECLARE res BOOLEAN;
BEGIN
IF _ids IS NULL then
RETURN false;
END IF;
EXECUTE format ( 'SELECT COUNT(*) = cardinality($1)
FROM %I.%I WHERE %I = ANY($2)',
_schema_name,_table_name,_column_name )
INTO res using _ids, _ids;
RETURN res;
END;
$$ language plpgsql stable security invoker;
Demo
Related
I need to replace schema and table name by parameters in following function (that is currently working perfectly):
CREATE OR REPLACE FUNCTION public.my_function_119()
RETURNS integer
LANGUAGE plpgsql
AS $function$
DECLARE _check INTEGER;
BEGIN
SELECT SUM("length"/1000)
FROM public."National_Grid_multiline"
INTO _check;
RETURN _check;
END
$function$
I have tried following solution (and its numerous variations) :
CREATE OR REPLACE FUNCTION public.my_function_119(schema text, tablename text)
RETURNS INTEGER
LANGUAGE plpgsql
AS
$function$
DECLARE _check INTEGER;
BEGIN
RETURN
'(SELECT SUM((length/1000))::integer FROM ' || schema || '."' || tablename || '")::integer INTO _check' ;
RETURN _check;
END
$function$
but keep running into following error code :
psycopg2.errors.InvalidTextRepresentation: invalid input syntax for type integer: "(SELECT SUM((length/1000))::integer FROM public."National_Grid_multiline")::integer INTO _check"
CONTEXT: PL/pgSQL function my_function_119(text,text) while casting return value to function's return type
Why is this not working ? The 'length' column contains float values.
You have to use dynamic SQL, because you cannot use a parameter for an identifier.
Also, make sure to avoid SQL injection by using format rather than concatenating strings:
EXECUTE
format(
'SELECT SUM((length/1000))::integer FROM %I.%I',
schema,
table_name
)
INTO _check';
You can try this :
CREATE OR REPLACE FUNCTION public.my_function_119(schema text, tablename text)
RETURNS INTEGER
LANGUAGE plpgsql
AS
$function$
DECLARE
res integer ;
BEGIN
EXECUTE E'
(SELECT SUM((length/1000))::integer INTO res FROM ' || schema || '."' || tablename || '"):: integer' ;
RETURN res ;
END ;
$function$
I have a bytea column in a table that contains a function decode(). What I have done to get the actual data is as follows:
select filename, convert_from(data,'UTF-8') from attachments limit 20; //this returns me decode function
select decode(E'...','hex'); // I am executing the above returned function
The above is fine as long as I have to select one row. But now my requirement is to get more than one result. How can I get the result in single query? I have tried using pl/pgsql
CREATE OR REPLACE FUNCTION get_data(integer, _type anyelement, OUT _result anyelement)
AS
$x$
BEGIN
EXECUTE
'SELECT ' || (select convert_from(data,'UTF-8') as data from attachments limit $1)
INTO _result;
END;
$x$
LANGUAGE plpgsql;
But this works only for single row and single column. What I want is a single query to fetch 2 columns without using pl/pgsql if possible. I am using this query from my Java based web app.
Thanks!
You need procedural code for this, since there is no provision for dynamic statements in SQL.
The following function converts all attachments:
CREATE FUNCTION getemall(
IN v_type anyelement,
OUT v_result anyelement
) RETURNS SETOF anyelement
LANGUAGE plpgsql AS
$$DECLARE
v_stmt text;
BEGIN
FOR v_stmt IN
SELECT convert_from(data,'UTF-8')
FROM attachments
LOOP
EXECUTE v_stmt INTO v_result;
RETURN NEXT;
END LOOP;
END;$$;
This is how I have written the function with few changes
CREATE OR REPLACE FUNCTION getmeall(tName text, fNameCol text, dataCol text,fSize
numeric)
RETURNS TABLE(bdata bytea, fname text) LANGUAGE plpgsql AS
$$DECLARE
v_stmt text;
v_name text;
BEGIN
FOR v_stmt,v_name IN
EXECUTE format('SELECT encode(%s, ''escape''), %s FROM %s
WHERE $1 IS NOT NULL AND $2 IS NOT NULL LIMIT $3'
, dataCol, fNameCol, tName)
USING dataCol, fNameCol, fSize
LOOP
fname:=v_name;
IF strpos(v_stmt,'decode') = 1 THEN
EXECUTE 'SELECT ' || v_stmt INTO bdata;
ELSE
bdata:=v_stmt;
END IF;
RETURN NEXT;
END LOOP;
END;$$;
And finally calling it this way.
select * from getmeall('attachments', '"filename"', '"data"',2)
CREATE OR REPLACE FUNCTION demo(vsql text, vals text[])
RETURNS void AS
$BODY$
BEGIN
execute vsql using vals;
END;
$BODY$
LANGUAGE plpgsql VOLATILE;
select demo('select $1 || $2', ARRAY['Barrett', ' Gilmour'] );
ERROR: there is no parameter $2
LINE 1: select $1 || $2
The error is that postgres does not understand that the two items in the array must be spread to the input parameters $1 and $2. It understand the entire array as value for $1
As discussed in SO question https://dba.stackexchange.com/questions/83278/postgressql-dynamic-execute-with-argument-values-in-array, the EXECUTE ... USING will treat an array as a single parameter.
You could try this hack, which edits the passed SQL statement so that $n becomes $1[n]
vsql := regexp_replace( vsql, '[$]([0-9]+)', '$1[\1]', 'g' );
The array vals is treated as a single parameter but the parameterised select can index into it to get the elements.
a_horse_with_no_name mentioned your other problem which is that your function doesn't actually return anything. If you want to see what is being executed:
-- anyelement is a polymorphic pseudo-type
-- problem with returning 'setof record' is that select * from so_demov3() gives error
-- ERROR: a column definition list is required for functions returning "record"
drop function so_demov3(anyelement, text, text[]);
create function so_demov3(rec_type anyelement, p_sql text, p_vals text[]) returns setof anyelement as
$$
declare v_sql text;
begin
-- edit the executable SQL stmt so that $n -> $1[n]
v_sql := regexp_replace( p_sql, '[$]([0-9]+)', '$1[\1]', 'g' );
return query execute v_sql using p_vals;
end;
$$
language plpgsql volatile;
And call it like this:
select * from so_demov3(null::text, 'select $1 || $2', array['Barrett', 'Gilmour'] );
Need Output from table with in clause in PostgreSQL
I tried to make loop or ids passed from my code. I did same to update the rows dynamically, but for select I m not getting values from DB
CREATE OR REPLACE FUNCTION dashboard.rspgetpendingdispatchbyaccountgroupidandbranchid(
IN accountgroupIdCol numeric(8,0),
IN branchidcol character varying
)
RETURNS void
AS
$$
DECLARE
ArrayText text[];
i int;
BEGIN
select string_to_array(branchidcol, ',') into ArrayText;
i := 1;
loop
if i > array_upper(ArrayText, 1) then
exit;
else
SELECT
pd.branchid,pd.totallr,pd.totalarticle,pd.totalweight,
pd.totalamount
FROM dashboard.pendingdispatch AS pd
WHERE
pd.accountgroupid = accountgroupIdCol AND pd.branchid IN(ArrayText[i]::numeric);
i := i + 1;
end if;
END LOOP;
END;
$$ LANGUAGE 'plpgsql' VOLATILE;
There is no need for a loop (or PL/pgSQL actually)
You can use the array directly in the query, e.g.:
where pd.branchid = any (string_to_array(branchidcol, ','));
But your function does not return anything, so obviously you won't get a result.
If you want to return the result of that SELECT query, you need to define the function as returns table (...) and then use return query - or even better make it a SQL function:
CREATE OR REPLACE FUNCTION dashboard.rspgetpendingdispatchbyaccountgroupidandbranchid(
IN accountgroupIdCol numeric(8,0),
IN branchidcol character varying )
RETURNS table(branchid integer, totallr integer, totalarticle integer, totalweight numeric, totalamount integer)
AS
$$
SELECT pd.branchid,pd.totallr,pd.totalarticle,pd.totalweight, pd.totalamount
FROM dashboard.pendingdispatch AS pd
WHERE pd.accountgroupid = accountgroupIdCol
AND pd.branchid = any (string_to_array(branchidcol, ',')::numeric[]);
$$
LANGUAGE sql
VOLATILE;
Note that I guessed the data types for the columns of the query based on their names. You have to adjust the line with returns table (...) to match the data types of the select columns.
I have some function on PostgreSQL 9.6 returning a cursor (refcursor):
CREATE OR REPLACE FUNCTION public.test_returning_cursor()
RETURNS refcursor
IMMUTABLE
LANGUAGE plpgsql
AS $$
DECLARE
_ref refcursor = 'test_returning_cursor_ref1';
BEGIN
OPEN _ref FOR
SELECT 'a' :: text AS col1
UNION
SELECT 'b'
UNION
SELECT 'c';
RETURN _ref;
END
$$;
I need to write another function in which a temp table is created and all data from this refcursor are inserted to it. But INSERT INTO ... FETCH ALL FROM ... seems to be impossible. Such function can't be compiled:
CREATE OR REPLACE FUNCTION public.test_insert_from_cursor()
RETURNS table(col1 text)
IMMUTABLE
LANGUAGE plpgsql
AS $$
BEGIN
CREATE TEMP TABLE _temptable (
col1 text
) ON COMMIT DROP;
INSERT INTO _temptable (col1)
FETCH ALL FROM "test_returning_cursor_ref1";
RETURN QUERY
SELECT col1
FROM _temptable;
END
$$;
I know that I can use:
FOR _rec IN
FETCH ALL FROM "test_returning_cursor_ref1"
LOOP
INSERT INTO ...
END LOOP;
But is there better way?
Unfortunately, INSERT and SELECT don't have access to cursors as a whole.
To avoid expensive single-row INSERT, you could have intermediary functions with RETURNS TABLE and return the cursor as table with RETURN QUERY. See:
Return a query from a function?
CREATE OR REPLACE FUNCTION f_cursor1_to_tbl()
RETURNS TABLE (col1 text) AS
$func$
BEGIN
-- MOVE BACKWARD ALL FROM test_returning_cursor_ref1; -- optional, see below
RETURN QUERY
FETCH ALL FROM test_returning_cursor_ref1;
END
$func$ LANGUAGE plpgsql; -- not IMMUTABLE
Then create the temporary table(s) directly like:
CREATE TEMP TABLE t1 ON COMMIT DROP
AS SELECT * FROM f_cursor1_to_tbl();
See:
Creating temporary tables in SQL
Still not very elegant, but much faster than single-row INSERT.
Note: Since the source is a cursor only the first call succeeds. Executing the function a second time would return an empty set. You would need a cursor with the SCROLL option and move to the start for repeated calls.
This function does INSERT INTO from refcursor. It is universal for all the tables. The only requirement is that all columns of table corresponds to columns of refcursor by types and order (not necessary by names).
to_json() does the trick to convert any primitive data types to string with double-quotes "", which are later replaced with ''.
CREATE OR REPLACE FUNCTION public.insert_into_from_refcursor(_table_name text, _ref refcursor)
RETURNS void
LANGUAGE plpgsql
AS $$
DECLARE
_sql text;
_sql_val text = '';
_row record;
_hasvalues boolean = FALSE;
BEGIN
LOOP --for each row
FETCH _ref INTO _row;
EXIT WHEN NOT found; --there are no rows more
_hasvalues = TRUE;
SELECT _sql_val || '
(' ||
STRING_AGG(val.value :: text, ',') ||
'),'
INTO _sql_val
FROM JSON_EACH(TO_JSON(_row)) val;
END LOOP;
_sql_val = REPLACE(_sql_val, '"', '''');
_sql_val = TRIM(TRAILING ',' FROM _sql_val);
_sql = '
INSERT INTO ' || _table_name || '
VALUES ' || _sql_val;
--RAISE NOTICE 'insert_into_from_refcursor(): SQL is: %', _sql;
IF _hasvalues THEN --to avoid error when trying to insert 0 values
EXECUTE (_sql);
END IF;
END;
$$;
Usage:
CREATE TABLE public.table1 (...);
PERFORM my_func_opening_refcursor();
PERFORM public.insert_into_from_refcursor('public.table1', 'name_of_refcursor_portal'::refcursor);
where my_func_opening_refcursor() contains
DECLARE
_ref refcursor = 'name_of_refcursor_portal';
OPEN _ref FOR
SELECT ...;