How to select duplicates by first order of appearance - sql

Question

I am looking to select unique values from a SQL database but I want to make sure that I am selecting only the first duplicate in order of appearance (in my case - date in the hospital, intime col)
You can see the code below.
I am trying to take only the IDs of the first time the patients were hospitalized which correspond to the "intime" col.
I have no absolute way to check that by ordering as I did and by using groupby, SQL will in fact return the id in the same order.
Thank you very much.
WITH ccupatients AS
(SELECT HADM_ID
FROM `physionet-data.mimiciii_clinical.icustays` i
WHERE first_careunit = 'CCU'
ORDER BY intime)
SELECT hadm_id
FROM ccupatients
GROUP BY hadm_id

Answer 1

Use ROW_NUMBER() if your RDBMS supports it: this works by ranking records by increasing intime within groups of records having the same ham_id, and then filtering in the outer query on the top record per group:
SELECT hadm_id
FROM (
SELECT hadm_id, ROW_NUMBER() OVER(PARTITION BY hadm_id ORDER BY intime) rn
FROM `physionet-data.mimiciii_clinical`.icustays
WHERE first_careunit = 'CCU'
) x
WHERE rn = 1
If you RDBMS does not support window functions such as ROW_NUMBER(), another option is to use a NOT EXISTS condition with a correlated subquery:
SELECT hadm_id
FROM `physionet-data.mimiciii_clinical`.icustays i
WHERE
first_careunit = 'CCU'
AND NOT EXISTS (
SELECT 1
FROM `physionet-data.mimiciii_clinical`.icustays i1
WHERE
i1.first_careunit = 'CCU'
AND i1.hadm_id = i.hadm_id
AND i1.intime < i.intime
)