json 格式的 InfluxDB 查询转换为带有 jq 的 csv，包括标签和字段

Question

I want to process data with a bash script but have trouble to get the InfluxDB output to the desired csv output with all tags and fields.

Below an example output from an influx query:

{
  "results": [
    {
      "series": [
        {
          "name": "tickerPrice",
          "tags": {
            "symbol": "AAVE",
            "symbolTo": "EUR"
          },
          "columns": [
            "time",
            "priceMean"
          ],
          "values": [
            [
              1614402874120627200,
              282.398263888889
            ]
          ]
        },
        {
          "name": "tickerPrice",
          "tags": {
            "symbol": "BTC",
            "symbolTo": "EUR"
          },
          "columns": [
            "time",
            "priceMean"
          ],
          "values": [
            [
              1614402874120627200,
              39189.756944444445
            ]
          ]
        }
      ]
    }
  ]
}

And I would like to transform it to:

"name","symbol","symbolTo","time","priceMean"
"tickerPrice","AAVE","EUR",1614402874120627200,282.398263888889
"tickerPrice","BTC","EUR",1614402874120627200,39189.756944444445

I have managed (google) to get the fields to a csv format but till now not managed to get all data in the csv. Here is the commands that I use for that:

$ jq -r '(.results[0].series[0].columns), (.results[0].series[].values[])'

Because this is not the only query I want to do it would be nice that it is universal for the content, so the number of fields and tags could be different.

Answer 1

The following produces the required output in a way that allows for multiple values of "time" in each .values array, but does not refer to the specific headers except for "name":

def headers:
 (.tags | keys_unsorted) as $tags
 | (["name"] + $tags + .columns); 

.results[0]
| (.series[0] | headers),
  (.series[] | ([.name, .tags[]] + .values[]))
| @csv

This of course assumes that the separate "series" are conformal.

Answer 2

Why you just don't specify csv format directly in influxdb CLI https://docs.influxdata.com/influxdb/v1.8/tools/shell/ :

-format 'json|csv|column' Specifies the format of the server responses.

So you won't need any result post processing.