Adding new tag to existing series


Is it possible to add, change, or remove tags on existing series without having to reinsert the series. I can’t seem to find any information about this in the docs.

@lrolsen It is not. We have some documentation on the tradeoffs we made when designing the database that might help.

Bummer :frowning:

What I am trying to achieve is to perform aggregation on any subset of the series.

In our use-case we have a large number of sensors scattered throughout Europe. Each sensor is represented by a series with a predefined number of tags (country, type etc.). Normally, the aggregations are based on these tags (all sensors in a country etc.) but from time to time we would like to be able to group the sensors differently and thus perform the aggregations on a grouping that can not necessarily be represented by the existing tag structure. As we have several years worth of data doing a rebuild of the series to alter the tag structure would be a big, expensive, and rather time consuming task. Also, as we can’t have any down time, this operation would most likely also be a rather tricky.

Does anybody have any suggestions as to how we can achieve this?

@lrolsen You can always do this client side and then rewrite the data back into the db for visualization. Kapacitor has some additional data manipulation tooling that can help, but for a case like this I would point you to our client libraries.

@jackzampolin Thanks, I will give it a go. I am a bit concerned about the amount of data though. It will typically require me to reinsert 50+% of the database.

Do you know whether there would be any chance of having this feature added if I were to add a feature request?. I would assume it would be a rather fairly common scenario when working with data from IOT devices. I work with data collected from smart meters and weather stations and typically need to slice and dice the data in a lot of different ways. With this limitation we are limited to only aggregating on the tag structure we devised when we initially collected the data (unless we reindex everything).

I’m also interested in this topic. It is very hard to plan every future changes. and it would be great to have features like to add tags to an existing series.
use case:
a company may only have 1 production line when it installs influxdb, so there’s no such tag called production line. but it may add a new production line. how to differentiate the new production line data if we can’t add a new tag?

@jackzampolin any update on this ? given influx is already planning 2.0 and flux.

I have similar use case.