Umami is a privacy-first web and product analytics platform. Traffic, campaigns, behavior, conversions, and revenue in one place. No cookies, no surveillance, self-hosted or in the cloud.
TypeScript analyticsaudience-segmentationchartscohort-analysisgoogle-analyticsproduct-analyticsstatisticsuser-journeyweb-analytics
import clickhouse from '@/lib/clickhouse';
import { EVENT_TYPE } from '@/lib/constants';
import { CLICKHOUSE, PRISMA, runQuery } from '@/lib/db';
import prisma from '@/lib/prisma';
import type { PropertyFilter, QueryFilters, TimeSeriesValue } from '@/lib/types';
const FUNCTION_NAME = 'getSessionDataNumericSeries';
export async function getSessionDataNumericSeries(
...args: [
websiteId: string,
propertyName: string,
metric: 'sum' | 'avg' | 'count',
filters: QueryFilters,
propertyFilters?: PropertyFilter[],
]
): Promise<TimeSeriesValue[]> {
return runQuery({
[PRISMA]: () => relationalQuery(...args),
[CLICKHOUSE]: () => clickhouseQuery(...args),
});
}
async function relationalQuery(
websiteId: string,
propertyName: string,
metric: 'sum' | 'avg' | 'count',
filters: QueryFilters,
propertyFilters: PropertyFilter[] = [],
) {
const { timezone = 'utc', unit = 'day' } = filters;
const { rawQuery, getDateSQL, parseFilters, getPropertyFilterQuery } = prisma;
const { filterQuery, cohortQuery, joinSessionQuery, queryParams } = parseFilters({
...filters,
websiteId,
timezone,
});
const { sql: pfSQL, params: pfParams } = getPropertyFilterQuery(
propertyFilters,
'session',
timezone,
);
const aggSql =
metric === 'avg'
? 'avg(cast(session_data.number_value as decimal))'
: metric === 'count'
? 'count(distinct session_data.session_id)'
: 'sum(cast(session_data.number_value as decimal))';
return rawQuery(
`
with filtered_sessions as (
select distinct website_event.session_id, website_event.website_id
from website_event
${cohortQuery}
${joinSessionQuery}
where website_event.website_id = {{websiteId::uuid}}
and website_event.created_at between {{startDate}} and {{endDate}}
and website_event.event_type != ${EVENT_TYPE.performance}
${filterQuery}
${pfSQL}
)
select
${getDateSQL('session_data.created_at', unit, timezone)} t,
${aggSql} y
from session_data
join filtered_sessions
on filtered_sessions.session_id = session_data.session_id
and filtered_sessions.website_id = session_data.website_id
where session_data.website_id = {{websiteId::uuid}}
and session_data.data_key = {{propertyName}}
and session_data.data_type = 2
group by 1
order by 1
`,
{ ...queryParams, propertyName, ...pfParams },
FUNCTION_NAME,
);
}
async function clickhouseQuery(
websiteId: string,
propertyName: string,
metric: 'sum' | 'avg' | 'count',
filters: QueryFilters,
propertyFilters: PropertyFilter[] = [],
): Promise<{ t: string; y: number }[]> {
const { timezone = 'UTC', unit = 'day' } = filters;
const { rawQuery, getDateSQL, parseFilters, getPropertyFilterQuery } = clickhouse;
const { filterQuery, cohortQuery, queryParams } = parseFilters({
...filters,
websiteId,
timezone,
});
const { sql: pfSQL, params: pfParams } = getPropertyFilterQuery(
propertyFilters,
'session',
timezone,
);
const aggSql =
metric === 'avg'
? 'avg(session_data.number_value)'
: metric === 'count'
? 'uniq(session_data.session_id)'
: 'sum(session_data.number_value)';
return rawQuery(
`
with filtered_sessions as (
select distinct website_event.session_id
from website_event
${cohortQuery}
where website_event.website_id = {websiteId:UUID}
and website_event.created_at between {startDate:DateTime64} and {endDate:DateTime64}
and website_event.event_type != ${EVENT_TYPE.performance}
${filterQuery}
${pfSQL}
)
select
${getDateSQL('session_data.created_at', unit, timezone)} as t,
${aggSql} as y
from session_data final
join filtered_sessions
on filtered_sessions.session_id = session_data.session_id
where session_data.website_id = {websiteId:UUID}
and session_data.data_key = {propertyName:String}
and session_data.data_type = 2
group by t
order by t
`,
{ ...queryParams, propertyName, ...pfParams },
FUNCTION_NAME,
);
}