Bigquery standard SQL: Filter out the duplicates while keeping the sequence of one column

Question

I'm writing query in BigQuery trying to export the distinct pages a session visited, following the sequence of when (PageVisit_time) the pages are visited (ascending), below is what I wrote and the outputs:

SELECT DISTINCT 
fullVisitorId||'.'||visitStartTime||'.'||visitNumber AS session_id, 
page.pagePath,
MIN(DATETIME_ADD(DATETIME(TIMESTAMP_SECONDS(visitStartTime),"America/New_York"), INTERVAL hits.time MILLISECOND))  AS PageVisit_time

from `xx.xx.ga_sessions_*`,
UNNEST(hits) AS hits
WHERE hits.type = "PAGE"
and date = '20220403' 
group by 1,2
order by 1, 3 desc

Outputs of the above query:

session_id	pagePath	PageVisit_time
123	/point	2022-04-03T11:26:13.719000
123	/point	2022-04-03T11:27:15.653820
123	/point-ad	2022-04-03T11:34:10.000000
123	/point-ad-next	2022-04-03T12:38:15.82340
123	/point	2022-04-03T12:50:18.123820

I want to keep only distinct session_id and pagePath in the final output, and I also want to make suer pagePath is listing follow the same sequence (ascending by PageVisit_time), any suggestions on how to make it work?

Ideal output:

session_id	pagePath
123	/point
123	/point-ad
123	/point-ad-next

Answer 1

If I'm not wrong, you want to get the first row for each pagePath in the session, and also the other attributes from these events.
If I'm correct, you can use array_agg function to get the first row you want. I'm simplifying the query here, so you can modify it based on your needs.

SELECT 
    fullVisitorId||'.'||visitStartTime||'.'||visitNumber AS session_id, 
    page.pagePath,
    array_agg(struct(
        visitStartTime,
        attribute1,
        attribute2,
        attribute
    ) order by visitStartTime limit 1) as attr
from `xx.xx.ga_sessions_*`,
UNNEST(hits) AS hits
WHERE hits.type = "PAGE"
    and date = '20220403' 
group by 1,2
order by session_id, attr.visitStartTime