I have a log table in which I keep user login records.

CREATE TABLE public.login
(
  id bigint NOT NULL,
  copy_id bigint,
  create_date timestamp without time zone,
  version integer NOT NULL DEFAULT 0,
  firm_id bigint,
  language_id bigint,
  user_id bigint,
  channel character varying(255),
  error character varying(255),
  ip character varying(255),
  username character varying(255),
  otp character varying(255),
  account_period_id bigint,
  login_code character varying(255),
  CONSTRAINT login_pkey PRIMARY KEY (id),
  CONSTRAINT fkilkkn9n0fkwoenwtpljjyhd89 FOREIGN KEY (user_id)
      REFERENCES public."user" (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fkj2ff261eu4uy2dc7ehpto877k FOREIGN KEY (language_id)
      REFERENCES public.language (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fkoi2rpag6j9pl6y21l74sxfic0 FOREIGN KEY (account_period_id)
      REFERENCES public.account_period (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fkpsi5sd9hd43btcd20brrbltf8 FOREIGN KEY (firm_id)
      REFERENCES public.firm (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION
)

For example, finding the number of logins of users but dividing them into weeks
1.2.3.4.5.6.7. I want to set it to week

• a user logged in 5 times in the 1st week

• a user 2 . logged in 3 times a week,

• a user logged in 5 times in the 3rd week

• a user 4 . logged in 3 times a week,

• a user logged in 5 times in the 5th week

• a user 6 . logged in 3 times a week,

• a user logged in 5 times in the 7th week

• b user logged in 51 times in the 1st week

• b user 2 . logged in 53 times in a week,

• b user logged in 54 times in the 3rd week

• b user 4 . logged in 39 times in a week,

• b user logged in 5 times in the 5th week

• b user 6 . logged in 13 times in a week,

• b user logged in 15 times in the 7th week

I tried the following:

SELECT
    user_id,
    DATE_TRUNC('week', create_date) AS week_start_date,
    COUNT(*) AS login_count
FROM
    login
GROUP BY
    user_id,
    DATE_TRUNC('week', create_date)
ORDER BY
    user_id,
    week_start_date;

output:
enter image description here

But that’s not what I want
What I want is to write zero even if the user with user_id is not logged in, but I could not achieve this.
enter image description here