Model alternative for jaeger_index table

[Etienne-Carriere]

On jaeger_index tables, the tags is coded as a nested array with key and values. It is good for the only usage of Jaeger-query but in our company we are using jaeger also for analytics purposes. Since Clickhouse 21.3, the Map type (https://clickhouse.com/docs/en/sql-reference/data-types/map/) is available. I think It could be a good alternative to Nested .

Do you have already made some performance (time and storage) tests with Map ? Could it be an acceptable contribution (with a flag to not activate it by default) ?

[levonet]

So do I. An alternative way is using tags.value[ indexOf(tags.key, 'some key') ]. I'm not saying that this is the best solution, but it works. As an example:

CREATE MATERIALIZED VIEW IF NOT EXISTS jaeger.adapter_search_materialized ON CLUSTER '{cluster}'
TO jaeger.adapter_search_local AS SELECT
    timestamp,
    resource,
    argUseCache,
    traceId,
    toDate(_tripDate) AS tripDate,
    tripDirection,
    tripSource,
    toUInt16OrZero(_tripCount) AS tripCount,
    tripSearchId,
    durationUs
FROM (
    SELECT
        timestamp,
        toInt32(tags.value[ indexOf(tags.key, 'resource.code') ]) AS resource,
        toBool(tags.value[ indexOf(tags.key, 'use_cache') ] != 'false') AS argUseCache,
        traceID AS traceId,
        splitByChar(',', tags.value[ indexOf(tags.key, 'trip.date') ]) AS _tripDate,
        splitByChar(',', tags.value[ indexOf(tags.key, 'trip.direction') ]) AS tripDirection,
        splitByChar(',', tags.value[ indexOf(tags.key, 'trip.source') ]) AS tripSource,
        splitByChar(',', tags.value[ indexOf(tags.key, 'trip.count') ]) AS _tripCount,
        tags.value[ indexOf(tags.key, 'trip.search_id') ] AS tripSearchId,
        durationUs
    FROM jaeger.jaeger_index_local
    WHERE operation = 'searchAndWaitResult'
) ARRAY JOIN _tripDate, tripDirection, tripSource, _tripCount