Graph message: "No data points" with subquery

[dmitrysdm]

grafana v6.1.6 clickhouse-datasource 1.8.1
Working Query: as result i see graph

SELECT
    $timeSeries as t,
    count() as c
FROM $table
WHERE $timeFilter
GROUP BY t
ORDER BY t

result in query inspector:

{
  "xhrStatus": "complete",
  "request": {
    "method": "GET",
    "url": "api/datasources/proxy/8/?query=SELECT%20%20%20%20%20(intDiv(toUInt32(EventTime)%2C%201)%20*%201)%20*%201000%20as%20t%2C%20%20%20%20%20count()%20as%20c%20FROM%20em.users%20WHERE%20EventDate%20%3E%3D%20toDate(1558019528)%20AND%20EventTime%20%3E%3D%20toDateTime(1558019528)%20GROUP%20BY%20t%20ORDER%20BY%20t%20FORMAT%20JSON"
  },
  "response": {
    "meta": [
      {
        "name": "t",
        "type": "UInt64"
      },
      {
        "name": "c",
        "type": "UInt64"
      }
    ],
    "data": [
      {
        "t": "1558019633000",
        "c": "91"
      },
      {
        "t": "1558019634000",
        "c": "216"
      }
    ],
    "rows": 2,
    "statistics": {
      "elapsed": 0.003194378,
      "rows_read": 9777,
      "bytes_read": 58662
    }
  }
}

But if i try to do subquery:

SELECT * FROM (
  SELECT
      $timeSeries as t,
      count() as c
  FROM $table
  WHERE $timeFilter
  GROUP BY t
  ORDER BY t
)
GROUP BY c, t

i see same result in query inspector as previously:

{
  "xhrStatus": "complete",
  "request": {
    "method": "GET",
    "url": "api/datasources/proxy/8/?query=SELECT%20*%20FROM%20(%20SELECT%20%20%20%20%20(intDiv(toUInt32(EventTime)%2C%201)%20*%201)%20*%201000%20as%20t%2C%20%20%20%20%20count()%20as%20c%20FROM%20em.users%20WHERE%20EventDate%20%3E%3D%20toDate(1558019589)%20AND%20EventTime%20%3E%3D%20toDateTime(1558019589)%20GROUP%20BY%20t%20ORDER%20BY%20t%20)%20GROUP%20BY%20c%2C%20t%20FORMAT%20JSON"
  },
  "response": {
    "meta": [
      {
        "name": "t",
        "type": "UInt64"
      },
      {
        "name": "c",
        "type": "UInt64"
      }
    ],
    "data": [
      {
        "t": "1558019634000",
        "c": "216"
      },
      {
        "t": "1558019633000",
        "c": "91"
      }
    ],
    "rows": 2,
    "statistics": {
      "elapsed": 0.003641224,
      "rows_read": 9777,
      "bytes_read": 58662
    }
  }
}

but instead graph, i see message:
"No data points"
without any messages about errors.

[dmitrysdm]

With grafana-6.2.0-beta2 i have the same result :(

[juris]

Same here with 6.7.1.

Moreover, it shows "No data" if I specify fixed time range. Presets like "Last X days" work fine. Defined ranges, like "From now-30 to now-1" produce "No data".

[juris]

Same here with 6.7.1.

Moreover, it shows "No data" if I specify fixed time range. Presets like "Last X days" work fine. Defined ranges, like "From now-30 to now-1" produce "No data".

Figured out why I'm seeing it. We have a DateTime column named "t". So I was able to overcome my issue by changing :

SELECT
    $timeSeries as t

to

SELECT
    $timeSeries as time

[Slach]

@dmitrysdm
if you change your query
from

SELECT * FROM (
  SELECT
      $timeSeries as t,
      count() as c
  FROM $table
  WHERE $timeFilter
  GROUP BY t
  ORDER BY t
)
GROUP BY c, t

to

SELECT * FROM (
  SELECT
      $timeSeries as t,
      count() as total
  FROM $table
  WHERE $timeFilter
  GROUP BY t
)
ORDER BY t

you will see results without "no data"

"no data" which you see related to behavior how to clickhouse-grafana plugin detect which columns should be represented as "groups" or "labels" and which columns should be represented as "value"
please look to:

• https://github.com/Vertamedia/clickhouse-grafana/blob/2.0.2/src/datasource.ts#L109
• https://github.com/Vertamedia/clickhouse-grafana/blob/2.0.2/src/datasource.ts#L164

when SQL query contains GROUP BY
all columns used in GROUP BY should be combined to "label values" or name of time series depends on how much columns not used in GROUP BY

so, when you use GROUP BY c, t you automatically exclude "c" from list of columns which should represent as "value"