Adding max id from a different table to a query in PostgreSQL - postgresql

I need to join the max(id) with is a bigint (and a primary key) from one table to another table. So, if the max id is 1234567890, then I would want the results to be:
name, max_id
'foo', 1234567890
'bar', 1234567890
'noo', 1234567890
'boo', 1234567890
'zoo', 1234567890
My first thought was just to write a function and market it as STABLE
CREATE OR REPLACE FUNCTION get_max_id() RETURNS bigint AS
$$
DECLARE
result bigint;
BEGIN
RAISE NOTICE 'entering get_max_id';
SELECT INTO result max(id) FROM mytable;
RETURN result;
END;
$$
LANGUAGE plpgsql STABLE;
Then I called it via a SQL statement like:
SELECT name, get_max_id() FROM a_really_big_table;
But, from my raise notice, it seems like this is getting called for each each record in a_really_big_table.
What is the best way to write this so that I can keep the PostgreSQL from doing a lookup on mytable for each row in a_really_big_table?
I'm using PostgreSQL 9.4.1


Could it be a solution?
=# SELECT name AS name,
(SELECT MAX(id) FROM your_table) AS max_id
FROM really_big_table;
name │ max_id
──────────────────┼─────────
44e5f8fc1f75fad2 │ 5307021
97fc0aa3a5e3da98 │ 5307021
ccb83afb938ad758 │ 5307021
523bf25914e5c9c0 │ 5307021
d2362badeb2eb7ea │ 5307021
b6e0b87df6d41523 │ 5307021
da754db1fa226453 │ 5307021
865d0e177cb0a3e8 │ 5307021
5904a07a273d8cba │ 5307021
952c75ef37d47dab │ 5307021
(10 rows)

Related

Convert Hexadecimals to timestamp in postgres

I have the following hexadecimal data that I want to convert using the postgres functions, the data they receive is the following:
#AVLData_TimeStamp #AVLData_GPSElement_Altitude
00000174C0FA7EA0 0140
And I want to convert them to the following data
#MOPO_FECHAHORA #MOPO_ALTITUD
2020-09-24 09:37:56.000 320
I have tried to do it using this function but have not succeeded
SELECT to_timestamp(#AVLData_TimeStamp)
I have two functions in SQLserver that do this but in postgres I don't know how to do it
First Function:
CREATE FUNCTION FUN_Unix_To_Date(
#FechaUnix CHAR(38)
)
RETURNS DATETIME
AS
BEGIN
DECLARE
#LocalTime BIGINT
,#Adjusted BIGINT
,#seconds BIGINT = #FechaUnix
SET #LocalTime = DATEDIFF(MS,GETDATE(),GETUTCDATE())
SET #Adjusted = #seconds - #LocalTime
RETURN (DATEADD(SECOND,CONVERT(BIGINT, LEFT(CAST(#Adjusted AS VARCHAR(100)),13))/1000, CAST('1970-01-01 00:00:00' AS DATETIME)))
END
Two Function:
CREATE FUNCTION HexadecimalToDec_v2(
#hexval CHAR(38)
)
RETURNS NUMERIC(38)
AS
BEGIN
DECLARE #i INT
,#digits INT
,#result NUMERIC
,#current_digit CHAR(1)
,#current_digit_dec NUMERIC
SET #digits = LEN(#hexval)
SET #i = 0
SET #result =0
WHILE #i <= #digits
BEGIN
SET #current_digit = SUBSTRING(#hexval, #i, 1)
IF #current_digit = 'A' OR
#current_digit = 'B' OR
#current_digit = 'C' OR
#current_digit = 'D' OR
#current_digit = 'E' OR
#current_digit = 'F'
SET #current_digit_dec = ASCII(#current_digit) - ASCII('A') + 10
ELSE
SET #current_digit_dec = CONVERT(INT,#current_digit)
SET #result = (#result * 16) + #current_digit_dec
SET #i = #i + 1
END
RETURN(#result)
END
and finally the function is used like this
#MOPO_FECHAHORA = (dbo.Unix_To_Date(dbo.FUN_HexadecimalToDec_v2(#AVLData_TimeStamp)))
And this is the result
#MOPO_FECHAHORA
2020-09-24 09:37:56.000
Thank you very much for your help.

Use the function from this answer to convert hexadecimal to decimal (you might want to use bigint rather than integer):
CREATE OR REPLACE FUNCTION hex_to_int(hexval text) RETURNS bigint
LANGUAGE plpgsql IMMUTABLE STRICT AS
$$DECLARE
result bigint;
BEGIN
EXECUTE 'SELECT x' || quote_literal(hexval) || '::bigint'
INTO result;
RETURN result;
END;$$;
The timestamp can then be converted with
to_timestamp(hex_to_int('00000174C0FA7EA0') / 1000.0)

Just to clarify the source of issue.
The problem is that the your value(s) is the "local" epoch but to_timestamp() function returns timestamp with timezone value. Lets try some example:
with t(x) as (values('2020-09-24 09:37:56'))
select
x::timestamp as srcts,
x::timestamptz as srctstz,
to_timestamp(extract(epoch from x::timestamp)) as cnvts,
to_timestamp(extract(epoch from x::timestamptz)) as cnvtstz
from t;
┌─[ RECORD 1 ]─────────────────────┐
│ srcts │ 2020-09-24 09:37:56 │
│ srctstz │ 2020-09-24 09:37:56+03 │
│ cnvts │ 2020-09-24 12:37:56+03 │ <- Here is the issue in our case
│ cnvtstz │ 2020-09-24 09:37:56+03 │
└─────────┴────────────────────────┘
As you can see the data source type is critical. That's why you got the increased (actually converted from UTC to the local time zone) value in the first try using #LaurenzAlbe answer.
To fix this issue you need to perform some "reverse" calculations:
with t(x) as (values('2020-09-24 09:37:56'))
select
x::timestamp as srcts,
x::timestamptz as srctstz,
to_timestamp(extract(epoch from x::timestamp)) as cnvts,
(to_timestamp(extract(epoch from x::timestamp)))::timestamptz at time zone 'utc' as cnvtsrecalc,
to_timestamp(extract(epoch from x::timestamptz)) as cnvtstz
from t;
┌─[ RECORD 1 ]┬────────────────────────┐
│ srcts │ 2020-09-24 09:37:56 │
│ srctstz │ 2020-09-24 09:37:56+03 │
│ cnvts │ 2020-09-24 12:37:56+03 │ <- Here is the issue in our case
│ cnvtsrecalc │ 2020-09-24 09:37:56 │ <- Issue fixed
│ cnvtstz │ 2020-09-24 09:37:56+03 │
└─────────────┴────────────────────────┘
The following function wraps all this logic including the conversion of the hex value to bigint:
create or replace function hex2ts(text)
returns timestamp
language sql
immutable
strict
as $$
select
(to_timestamp(('x' || lpad($1, 16, '0'))::bit(64)::bigint / 1000.0))::timestamptz at time zone 'utc'
$$;

What is the data type of column_name in information_schema.columns?

What is the data type of column_name in information_schema.columns?
i tried this but didnt get
code:
CREATE OR REPLACE FUNCTION table_name_sunitest118(
tbl character varying)
RETURNS TABLE(column_name text)
LANGUAGE 'plpgsql'
AS $BODY$
BEGIN
RETURN QUERY
EXECUTE FORMAT('SELECT column_name FROM information_schema.columns where table_name=$1')
using tbl;
END
$BODY$;

In general you can use pg_typeof() to get that information:
select pg_typeof(column_name) from information_schema.columns limit 1;
┌───────────────────────────────────┐
│ pg_typeof │
├───────────────────────────────────┤
│ information_schema.sql_identifier │
└───────────────────────────────────┘
(1 row)
The type is sql_identifier, and you can cast it to text with column_name::text in your function.
CREATE OR REPLACE FUNCTION table_name_sunitest118(
tbl character varying)
RETURNS TABLE(column_name text)
LANGUAGE 'plpgsql'
AS $BODY$
BEGIN
RETURN QUERY
EXECUTE FORMAT(
'SELECT column_name::text FROM information_schema.columns where table_name=$1'
) using tbl;
END
$BODY$;
CREATE FUNCTION
select * from table_name_sunitest118('users');
┌─────────────┐
│ column_name │
├─────────────┤
│ email │
│ pass │
│ role │
│ id │
│ email │
└─────────────┘
(5 rows)

In MySQL, describe information_schema.columns says varchar(64).
The Postgres docs for the columns table says it is a sql_identifier defined as...
A character string. This type is used for SQL identifiers, the type character_data is used for any other kind of text data.

In psql you can use \gdesc command. It shows type on client side.
postgres=# select column_name from information_schema.columns \gdesc
┌─────────────┬──────┐
│ Column │ Type │
╞═════════════╪══════╡
│ column_name │ name │
└─────────────┴──────┘
(1 row)
pg_typeof function shows type on server side
postgres=# select pg_typeof(column_name) from information_schema.columns limit 1;
┌───────────────────────────────────┐
│ pg_typeof │
╞═══════════════════════════════════╡
│ information_schema.sql_identifier │
└───────────────────────────────────┘
(1 row)

How to get executed client details inside PL/pgSQL function

could you please let me know, how can i get the function executed client(single client only) details inside the PL/pgSQL function(something $whoexecuted). i'm getting the current all active client details.
as per the following code.
begin
select ps.usename, ps.client_addr,ps.application_name from pg_stat_activity ps
where state = 'active'
and ps.usename in('user1') into pg_stat_activity_rowtype
;

The content of pg_stat_activity is available from PLpgSQL too. So you can use
do $$
declare r record;
begin
for r in select * from pg_stat_activity where pid = pg_backend_pid()
loop
raise notice '%', r;
end loop;
end;
$$;
These data can be collected from some source:
select current_user, session_user, current_database(), inet_client_addr(), current_setting('application_name');
┌──────────────┬──────────────┬──────────────────┬──────────────────┬─────────────────┐
│ current_user │ session_user │ current_database │ inet_client_addr │ current_setting │
╞══════════════╪══════════════╪══════════════════╪══════════════════╪═════════════════╡
│ pavel │ pavel │ postgres │ │ psql │
└──────────────┴──────────────┴──────────────────┴──────────────────┴─────────────────┘
(1 row)
More in documentation:
https://www.postgresql.org/docs/current/static/functions-admin.html
https://www.postgresql.org/docs/current/static/functions-info.html

Simple string return function with default param passed as NULL, returns NULL instead of string

I have the following function:
CREATE OR REPLACE FUNCTION public.get_string(cmd_type text, udf_name text,
group_name character varying DEFAULT 'usage'::character varying)
RETURNS text
LANGUAGE plpgsql
AS $function$
BEGIN
return 'This is the string: '''|| group_name ||''''::text;
END;
$function$
When i call it like this:
select public.get_string('test', 'myudf!', group_name=>null::character varying);
It returns NULL.
I expect it to at least return:
This is the string: ''
However, when I call it like this:
select public.get_string('test', 'myudf!');
The I get the expected:
This is the string: 'usage'
Why does passing NULL to an optional parameter make the entire string NULL?

It is not mystic - any operation over NULL value is NULL again.
postgres=# select ('Hello' || null) is null ;
┌──────────┐
│ ?column? │
╞══════════╡
│ t │
└──────────┘
(1 row)
You should to use a coalesce function and sanitize expression against NULL value.
postgres=# select ('Hello' || coalesce(null,'')) ;
┌──────────┐
│ ?column? │
╞══════════╡
│ Hello │
└──────────┘
(1 row)
Maybe you know a Oracle database, where NULL and empty strings are equal. But it is true only for Oracle, elsewhere NULL is NULL and it is more aggressive.

Postgres overlap arrays from one column

I have a Table "A" with one column "col1" where each record is a array of integers.
col1
-----
{1,2,3,4}
{1,2,6,7}
{1,2,3,8,9}
I like to have one row as result which contains the overlap or intersect of all arrays in "col1".
select overlap(col1) from A;
result
-----
{1,2}

You should to define custom aggregate for this purpose:
CREATE OR REPLACE FUNCTION public.overlap_array_aggregate(anyarray, anyarray)
RETURNS anyarray
LANGUAGE plpgsql STRICT
AS $function$
BEGIN
RETURN ARRAY(SELECT unnest($1) INTERSECT SELECT unnest($2));
END;
$function$
CREATE AGGREGATE array_overlap_agg (
basetype = anyarray,
sfunc = overlap_array_aggregate,
stype = anyarray );
Then it is working as you expect:
postgres=# SELECT * FROM foo;
┌─────────────┐
│ a │
╞═════════════╡
│ {1,2,3,4} │
│ {1,2,6,7} │
│ {1,2,3,8,9} │
└─────────────┘
(3 rows)
postgres=# SELECT array_overlap_agg(a) FROM foo;
┌───────────────────┐
│ array_overlap_agg │
╞═══════════════════╡
│ {1,2} │
└───────────────────┘
(1 row)