[influxdb] Down-sampling best practice

[o***@gmail.com]

Hi,

I have managed to down-sample 10 second inputs to 2 minute inputs with the MEAN(*) feature.

(see example 3 in: https://docs.influxdata.com/influxdb/v1.5/query_language/continuous_queries/ )

The problem is that only the 'value' column has been down-sampled. the other columns are empty.

example:
original rp has:

| time | host | servername | type | value |
|---------|---------|-------------|----------------------|-------|
| 3:00:00 | server1 | server1:web | Performance counters | 3 |
| 3:00:00 | server2 | server2:web | Performance counters | 5 |
| 3:00:10 | server1 | server1:web | Performance counters | 7 |
| 3:00:10 | server2 | server2:web | Performance counters | 8 |

the downsampled rp has:

| time | host | servername | type | mean_value |
|---------|---------|-------------|----------------------|------------|
| 3:00:00 | | | | 7 |
| 3:00:00 | | | | 9 |


when quering the data from the grafana for example everything breaks because it depends on the other columns as well.
I know there is no MEAN() for string, but how can I achieve what I need?

p.s. the command was:
SELECT MEAN(*) INTO "database"."other".:MEASUREMENT FROM database.autogen./.*/ GROUP BY time(2m)
And later on I'll do it automatically with a Continuous Query.

Thanks in advanced.