Rails query join to internal table count - ruby-on-rails

Question

I have two tables:
Player --> has_many FantasyStarts
FantasyStarts --> belongs_to Player
In english, I am trying to find all SeasonStats from the last year and week that were the first entry for that player in the whole table (ie who started last week and it was their first 'start' ever recorded in the table). I have constructed a bare sql query that works great, but I am trying to improve my rails query skills:
FantasyStart.includes(:player).find_by_sql("with count_table as (select player_id, count(*) as num_starts from fantasy_starts where position != 'BN' group by player_id) select * from fantasy_starts join count_table on fantasy_starts.player_id = count_table.player_id where week = 13 and year = 2019 and count_table.num_starts = 1 and position != 'BN' ")
I have also figure out a 'rails way' of creating the map of counts I would join to:
FantasyStart.where.not(position: 'BN').group(:player_id).count
which yields:
=> {184=>2, 2562721=>5, 2540215=>12, 100004=>57, 100001=>26, 100006=>62, 2505785=>5, 2561029=>1, 2541316=>1, 2558954=>1, 2552408=>12, 2532820=>60, 2507999=>25, 2506194=>16, 2505600=>18, 2532977=>6, 2507164=>75, 2495441=>4, 100022=>23, 2543704=>2, 2532807=>1, 81288=>1, 2550658=>4, 2506386=>15, 2560809=>11, 2533349=>4, 2560735=>3, 2557976=>19, 2556521=>15,....
hypothetically i could use the map above to filter the main query and only include player_ids that have a count of 1 in that map.
any good ideas for a next step?

Answer 1

Answering to your answer;
That's because you're chaining pluck to your "nested" query, that way ActiveRecord isn't able to perform a subquery, and instead of that it makes another query and use the result of that, which is an array to use with an IN clause.
Try instead removing the pluck, leaving just select:
FantasyStart.where(week: 13, year: 2019).where.not(position: 'BN').where(player_id: FantasyStart.where.not(position: 'BN').group(:player_id).having('count(*) = 1').select(:player_id))
It should generate something like this:
SELECT fantasy_starts.*
FROM fantasy_starts
WHERE fantasy_starts.week = $1
AND fantasy_starts.year = $2
AND fantasy_starts.position != $3
AND fantasy_starts.player_id IN (
SELECT fantasy_starts.player_id
FROM fantasy_starts
WHERE fantasy_starts.position != $4
GROUP BY fantasy_starts.player_id
HAVING (count(*) = 1)
) [[week, 13], [year, 2019], [position, BN], [position, BN]]
It seems you could shorten the query by removing the != from the subquery:
FantasyStart
.where.not(position: 'BN')
.where(player_id: FantasyStart
.group(:player_id)
.having('count(*) = 1')
.select(:player_id), week: 13, year: 2019)

Answer 2

hey i think i made progress curious what one of you more experience folk might think:
FantasyStart.where(week: 13, year: 2019).where.not(position: 'BN').where(player_id: FantasyStart.select(:player_id).where.not(position: 'BN').group(:player_id).having('count(*) = 1').pluck(:player_id))
this generates:
(8.4ms) SELECT "fantasy_starts"."player_id" FROM "fantasy_starts" WHERE "fantasy_starts"."position" != $1 GROUP BY "fantasy_starts"."player_id" HAVING (count(*) = 1) [["position", "BN"]]
FantasyStart Load (2.4ms) SELECT "fantasy_starts".* FROM "fantasy_starts" WHERE "fantasy_starts"."week" = $1 AND "fantasy_starts"."year" = $2 AND "fantasy_starts"."position" != $3 AND "fantasy_starts"."player_id" IN ($4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18, $19, $20, $21, $22, $23, $24, $25, $26, $27, $28, $29, $30, $31, $32, $33, $34, $35, $36, $37, $38, $39, $40, $41, $42, $43, $44, $45, $46, $47, $48, $49, $50, $51, $52, $53, $54, $55, $56, $57, $58, $59, $60, $61, $62, $63, $64, $65, $66, $67, $68, $69, $70, $71, $72, $73, $74, $75, $76, $77, $78, $79, $80, $81, $82, $83, $84, $85, $86, $87, $88, $89, $90, $91, $92, $93, $94, $95, $96, $97, $98, $99, $100, $101, $102, $103, $104, $105, $106, $107, $108, $109, $110, $111, $112, $113, $114, $115, $116, $117, $118, $119) LIMIT $120
this is still two queries so i know it could be better...