fix(queries): count custom events correctly on the hourly rollup, trying to address greptile findings

`website_event_stats_hourly` sets `views` to
`sumIf(1, event_type NOT IN (2, 5))`, so `views` is 0 on custom event
rows and the `events_count` join was summing zeroes. Switched it to
`sum(length(event_name))`, which is how `getWebsiteSessionStats`
already counts custom events on that table.

Also dropped the `getEventExpandedMetrics` change from the previous
commit. That report is filtered to custom events, so every visit it
lists has already fired at least one, and its `bounces` column counts
visits where the event fired exactly once rather than visits with a
single pageview. The threshold rule does not apply there, so the query
is left untouched.
This commit is contained in:
Federico D'Eredità committed 2026-07-26 18:49:59 +02:00
1 parent 82a5a12208
commit 92e9b276d2
2 files changed
+7 -24

No files matched your search

@@ -1,5 +1,5 @@
import clickhouse from '@/lib/clickhouse';
import { BOUNCE_THRESHOLD, EVENT_TYPE, FILTER_COLUMNS, SESSION_COLUMNS } from '@/lib/constants';
import { EVENT_TYPE, FILTER_COLUMNS, SESSION_COLUMNS } from '@/lib/constants';
import { CLICKHOUSE, PRISMA, runQuery } from '@/lib/db';
import prisma from '@/lib/prisma';
import type { QueryFilters } from '@/lib/types';
@@ -54,7 +54,7 @@ async function relationalQuery(
sum(t.c) as "pageviews",
count(distinct t.session_id) as "visitors",
count(distinct t.visit_id) as "visits",
sum(case when t.c = 1 and (t.events_count - t.c) < ${BOUNCE_THRESHOLD} then 1 else 0 end) as "bounces",
sum(case when t.c = 1 then 1 else 0 end) as "bounces",
sum(${getTimestampDiffSQL('t.min_time', 't.max_time')}) as "totaltime"
from (
select
@@ -63,19 +63,10 @@ async function relationalQuery(
website_event.visit_id,
count(*) as "c",
min(website_event.created_at) as "min_time",
max(website_event.created_at) as "max_time",
max((
select count(*)
from website_event we2
where we2.website_id = website_event.website_id
and we2.session_id = website_event.session_id
and we2.visit_id = website_event.visit_id
and we2.created_at between {{startDate}} and {{endDate}}
and we2.event_type = ${EVENT_TYPE.customEvent}
)) as "events_count"
max(website_event.created_at) as "max_time"
from website_event
${cohortQuery}
${joinSessionQuery}
${joinSessionQuery}
where website_event.website_id = {{websiteId::uuid}}
and website_event.created_at between {{startDate}} and {{endDate}}
${filterQuery}
@@ -113,7 +104,7 @@ async function clickhouseQuery(
sum(t.c) as "pageviews",
uniq(t.session_id) as "visitors",
uniq(t.visit_id) as "visits",
sumIf(1, t.c = 1 and (toInt64(ifNull(e.events_count, 0)) - toInt64(t.c)) < ${BOUNCE_THRESHOLD}) as "bounces",
sum(if(t.c = 1, 1, 0)) as "bounces",
sum(max_time-min_time) as "totaltime"
from (
select
@@ -131,15 +122,7 @@ async function clickhouseQuery(
${filterQuery}
group by name, session_id, visit_id
) as t
left join (
select session_id, visit_id, toUInt32(count()) as events_count
from website_event
where website_id = {websiteId:UUID}
and created_at between {startDate:DateTime64} and {endDate:DateTime64}
and event_type = ${EVENT_TYPE.customEvent}
group by session_id, visit_id
) as e using (session_id, visit_id)
group by name
group by name
order by visitors desc, visits desc
limit ${limit}
offset ${offset}
+1 -1
View File
@@ -153,7 +153,7 @@ async function clickhouseQuery(
group by session_id, visit_id
) as t
left join (
select session_id, visit_id, toUInt32(sumIf(views, event_type = ${EVENT_TYPE.customEvent})) as events_count
select session_id, visit_id, toUInt32(sum(length(event_name))) as events_count
from website_event_stats_hourly
where website_id = {websiteId:UUID}
and created_at between {startDate:DateTime64} and {endDate:DateTime64}