RegExp for custom dimension

[SCadilhac]

Hello,

Thanks for the great work on this project.

It would be great to be able to derive custom dimensions from other fields (e.g. interface description) using regular expressions.
From my understanding, this sits in between two existing mechanisms:

• Classifiers - Expr-based rules, executed by the ingestion code; can use RegExp, but only apply to predefined dimensions (Connectivity, Provider).
• Custom dictionaries - Executed by Clickhouse upon data insertion, custom configured dimensions, but implement only CSV lookup at the moment.

What would be the right direction to go?

[vincentbernat]

I would check if the appropriate functionality is present in ClickHouse. I don't think this is possible, but maybe this is and this will make everything easier.

[vincentbernat]

@margau any advice?

[margau]

Unfortunately, I have no proper dev setup available, therefore can't have a practical look at this right now.

It seems likely possible with CustomDicts (however, CustomDicts should be renamed then, because a RegExp is no dict lookup).

I could imagine some kind of switch between e.g. "DictLookup" and "Regex" source type in https://github.com/akvorado/akvorado/blob/main/common/schema/root.go#L91 with the later doing something e.g. with https://clickhouse.com/docs/en/sql-reference/functions/string-replace-functions