Performance issues on a join - performance

Question

I am joining two tables with bad performances.
Table1 :
Period
Zone
Country
company
product
01/01/2020
EMEA
DE
WKDM2
Product1
01/02/2020
EMEA
DE
PRL56
Product1
01/03/2020
EMEA
UK
ORD56
Product2
01/04/2020
EMEA
DE
GFDS
Product3
01/05/2020
EMEA
FR
24GFDSGF2
Product1
01/06/2020
EMEA
DE
2GFSDG37
Product3
01/07/2020
EMEA
IT
2GFDSG35
Product1
01/08/2020
EMEA
DE
23GSFDG6
Product4
01/09/2020
EMEA
DE
23GSFDG5
Product6
01/10/2020
EMEA
IT
24GSFD1
Product1
01/11/2020
EMEA
DE
23GSDF6
Product3
01/12/2020
EMEA
FI
24GFSDG1
Product8
Table2:
Period
Zone
Country
Quarter
Year
company
product
01/01/2020
EMEA
DE
01/01/2020
01/01/2020
WKDM2
Product1
01/02/2020
EMEA
DE
01/01/2020
01/01/2020
PRL56
Product1
01/03/2020
EMEA
UK
01/01/2020
01/01/2020
ORD56
Product2
01/04/2020
EMEA
DE
01/04/2020
01/01/2020
GFDS
Product3
01/05/2020
EMEA
FR
01/04/2020
01/01/2020
24GFDSGF2
Product1
01/06/2020
EMEA
DE
01/04/2020
01/01/2020
2GFSDG37
Product3
01/07/2020
EMEA
IT
01/07/2020
01/01/2020
2GFDSG35
Product1
01/08/2020
EMEA
DE
01/07/2020
01/01/2020
23GSFDG6
Product4
01/09/2020
EMEA
DE
01/07/2020
01/01/2020
23GSFDG5
Product6
01/10/2020
EMEA
IT
01/10/2020
01/01/2020
24GSFD1
Product1
01/11/2020
EMEA
DE
01/10/2020
01/01/2020
23GSDF6
Product3
01/12/2020
EMEA
FI
01/10/2020
01/01/2020
24GFSDG1
Product8
In my exemple, the data is the same but in Production ENV, Company on table 1 is the source of truth. And product on 2nd table is the source of truth.
Table1 have 6M rows and table2 600K rows.
When i join like this i have poor performances, how can i improve this ?:
SELECT R."Period",R."Zone",R."Country",S."Quarter",S."Year",R."Company",S."Product"
FROM table1 AS R,
table2 AS S
WHERE R."Period" = S."Period"
AND R."Zone" = S."Zone"
AND R."Country"=S."Country"
GROUP BY 1,2,3,4,5,6,7
UPDATE:
TEST DATASET :
CREATE OR REPLACE TEMPORARY TABLE "TMP_TEST1" (
"Period" TIMESTAMP,
"Country" VARCHAR,
"Quarter" TIMESTAMP,
"Year" TIMESTAMP,
"Company" VARCHAR,
"Product" VARCHAR
);
INSERT INTO "TMP_TEST1"
VALUES
('01/01/2020','DE','01/01/2020 ','01/01/2020 ','WKDM2 ','Product1'),
('01/01/2020','DE','01/01/2020 ','01/01/2020 ','2GFSDG37 ','Product1'),
('01/02/2020','DE','01/01/2020 ','01/01/2020 ','ORD56 ','Product2'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','GFDS ','Product3'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','24GFDSGF2 ','Product1'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','24GSFD1 ','Product1'),
('01/04/2020','DE','01/04/2020 ','01/01/2020 ','2GFSDG37 ','Product4'),
('01/04/2020','DE','01/04/2020 ','01/01/2020 ','23GSFDG5 ','Product6'),
('01/05/2020','DE','01/04/2020 ','01/01/2020 ','23GSDF6 ','Product3'),
('01/06/2020','DE','01/04/2020 ','01/01/2020 ','24GSFD1 ','Product8');
CREATE OR REPLACE TEMPORARY TABLE "TMP_TEST2" (
"Period" TIMESTAMP,
"Country" VARCHAR,
"Quarter" TIMESTAMP,
"Year" TIMESTAMP,
"Company" VARCHAR,
"Product" VARCHAR
);
INSERT INTO "TMP_TEST2"
VALUES
('01/01/2020','DE','01/01/2020 ','01/01/2020 ','WKDM2 ','Product1'),
('01/01/2020','DE','01/01/2020 ','01/01/2020 ','2GFSDG37 ','Product1'),
('01/02/2020','DE','01/01/2020 ','01/01/2020 ','ORD56 ','Product2'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','GFDS ','Product3'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','24GFDSGF2 ','Product1'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','2GFSDG37 ','Product3'),
('01/03/2020','DE','01/01/2020 ','01/01/2020 ','24GSFD1 ','Product1'),
('01/04/2020','DE','01/04/2020 ','01/01/2020 ','2GFSDG37 ','Product4'),
('01/04/2020','DE','01/04/2020 ','01/01/2020 ','23GSFDG5 ','Product6'),
('01/04/2020','DE','01/04/2020 ','01/01/2020 ','24GSFD1 ','Product1'),
('01/05/2020','DE','01/04/2020 ','01/01/2020 ','23GSDF6 ','Product3'),
('01/05/2020','DE','01/04/2020 ','01/01/2020 ','23GSDF6 ','Product9'),
('01/06/2020','DE','01/04/2020 ','01/01/2020 ','24GSFD1 ','Product8');
QUERY:
SELECT DISTINCT T1."Period",
T1."Country",
T1."Quarter",
T1."Year",
T1."Company",
T2."Product"
FROM TMP_TEST1 AS T1 INNER JOIN TMP_TEST2 AS T2
ON T1."Period" = T2."Period"
AND T1."Country"=T2."Country"
GROUP BY 1,2,3,4,5,6
With this test dataset you will see that i loose Products when there is no company. I don't know how to break this relation. I hope i am clear enough for you.

Answer 1

Firstly, given the data in the 2 source tables you have shown, please also provide the result you want to see.
For the query you have given, the correct way to write it, using ANSI SQL, is as follows:
SELECT R."Period",R."Zone",R."Country",S."Quarter",S."Year",R."Company",S."Product"
FROM table1 AS R
INNER JOIN table2 AS S ON
R."Period" = S."Period"
AND R."Zone" = S."Zone"
AND R."Country"=S."Country"
GROUP BY 1,2,3,4,5,6,7
To repeat my questions (and add some more)
How long is the query currently taking?
How long, roughly, should it take for you to consider the performance to be acceptable?
Why are you using GROUP BY when you have no aggregate functions in your query? If you want a distinct list (and your query is definitely producing duplicates) then use SELECT DISTINCT...
What do you mean by "cartesian calculations"? Do you mean cartesian joins and, if you do, why have you mentioned them as you don't have a cartesian join?
Response to Comments
ANSI SQL joins are much easier to read (and debug) as all the join information is in the JOIN statements and all the filtering conditions are in the WHERE statements - plus it is the industry standard so it would be a good idea for you, as a beginner, to get used to using it now rather than learning bad practices. Imagine if you were joining 20 tables with a mixture of inner/outer/left/right joins - the syntax you are using would be pretty incomprehensible whereas ANSI SQL join syntax would be simple to understand.
You still haven't provided the output that you are expecting to see, based on your source tables - so anyone trying to help you is left guessing what it is you are trying to achieve.
You also haven't provided the Explain Plan so no-one can see how your query is executing and therefore what the problem might be. In Snowflake, go to History, click on the relevant Query ID, click on Profile and then attach a screenshot showing all the steps being run, the execution time, the statistics, etc.

Answer 2

So I started making some tables that have 6M rows of data and 600K to show how joins via text is bad and use id's
CREATE TABLE table1 AS
with periods AS (
SELECT dateadd('day', SEQ8(), '1999-01-01'::date) as period
FROM TABLE(GENERATOR(rowcount=>1000))
), zones AS (
SELECT column1 as zone
,seq8() as zone_id
FROM VALUES ('EMEA')
), countries AS (
SELECT column1 as country
,seq8() as country_id
FROM VALUES ('DE'),('UK'),('FR'),('IT'),('NZ'),('AU'),('US'),('CA'),('xx'),('yy')
), company AS (
SELECT seq8() as comp_id
,hash(comp_id) as h_comp_id
,h_comp_id::text as t_h_comp_id
FROM TABLE(GENERATOR(rowcount=>600))
)
SELECT p.*, z.*, c.*, co.*
FROM periods p
JOIN zones z ON true
JOIN countries c ON true
JOIN company co ON true
;
CREATE TABLE table2 AS
SELECT *, YEAR(period) as year, QUARTER(period) as quarter
FROM table1
LIMIT 600000;
and then ran the SQL
SELECT R.Period,R.Zone,R.Country,S.Quarter,S.Year,R.comp_id
FROM table1 AS R
JOIN table2 AS S
ON R.Period = S.Period
AND R.Zone = S.Zone
AND R.Country=S.Country;
-- 1m27s
And it took ages, so ran my "fast" SQL
SELECT R.Period,R.Zone,R.Country,S.Quarter,S.Year,R.comp_id
FROM table1 AS R
JOIN table2 AS S
ON R.Period = S.Period
AND R.Zone_id = S.Zone_id
AND R.Country_id=S.Country_id;
-- 1m18s
and it took ages also.
Looking at the profile 90% of the time was getting the results.
SELECT S.Year, count(*) as c
FROM table1 AS R
JOIN table2 AS S
ON R.Period = S.Period
AND R.Zone = S.Zone
AND R.Country=S.Country
GROUP BY 1;
-- 9s
swapping to an aggregate to avoid the fetch, bad joins are 9 seconds
SELECT S.Year, count(*) as c
FROM table1 AS R
JOIN table2 AS S
ON R.Period = S.Period
AND R.Zone_id = S.Zone_id
AND R.Country_id=S.Country_id
GROUP BY 1;
-- 6s
and good joins are 6 seconds. So text is "still bad" but really fetching 3.6 millions rows of text is really slow.