Time series

Some APIs return parallel arrays of metric values (one entry per day in the requested range) instead of per-date objects. The time_series adapter expands those arrays into one row per date, using the request's start and end times to assign each index its calendar date.

Below you can find a before/after example. The original response holds metric arrays indexed by day:

{
  "data": [
    {
      "id": "1647935464530235393",
      "id_data": [
        {
          "metrics": {
            "impressions": [
              0,
              327,
              55
            ],
            "app_clicks": [
              0,
              6,
              2
            ]
          }
        }
      ]
    }
  ],
  "request": {
    "params": {
      "start_time": "2023-03-31T21:00:00Z",
      "end_time": "2023-04-02T21:00:00Z",
      "granularity": "DAY"
    }
  }
}

After the adapter runs, each per-date entry becomes its own object inside id_data, with the date attached:

{
  "data": [
    {
      "id": "1647935464530235393",
      "id_data": [
        {
          "date": "2023-03-31",
          "impressions": 0,
          "app_clicks": 0
        },
        {
          "date": "2023-04-01",
          "impressions": 327,
          "app_clicks": 6
        },
        {
          "date": "2023-04-02",
          "impressions": 55,
          "app_clicks": 2
        }
      ]
    }
  ]
}

Configuration:

"response": {
  "dataRows": {
    "source": "jsonPath",
    "value": "$.data.*.id_data.*"
  },
  "dataAdapters": {
    "items": [
      {
        "type": "time_series",
        "config": {
          "dataPath": "$.data.*",
          "valuesPath": "$.id_data.0.metrics",
          "startTimePath": "$.request.params.start_time",
          "endTimePath": "$.request.params.end_time",
          "granularity": {
            "source": "jsonPath",
            "value": "$.request.params.granularity"
          },
          "dataDestinationPath": "$.id_data",
          "keyName": "date"
        }
      }
    ]
  }
}

Below you can find example field definitions that read from the transformed structure:

[
  {
    "id": "date",
    "label": "Date",
    "dataType": "string.time.date",
    "value": {
      "source": "jsonPath",
      "value": "$.date"
    }
  },
  {
    "id": "impressions",
    "label": "Impressions",
    "dataType": "int.number.value",
    "value": {
      "source": "jsonPath",
      "value": "$.impressions"
    }
  },
  {
    "id": "app_clicks",
    "label": "App clicks",
    "dataType": "int.number.value",
    "value": {
      "source": "jsonPath",
      "value": "$.app_clicks"
    }
  }
]

Combining with flatten_values

With time_series alone, each row holds only the per-date fields — the parent's other fields (such as id) cannot be accessed easily on those rows. If your report needs fields from both the parent level (for example to know which campaign or entity each per-date row belongs to) and the per-date values, chain time_series with the Flatten values adapter so each per-date entry becomes its own row at $.data.* that also carries the parent's other fields. The per-date keys are prefixed with the source array's name (for example id_data_date, id_data_impressions).

After both adapters run, each per-date entry becomes its own row under data, with the parent's id preserved and the per-date metric keys prefixed with id_data_:

{
  "data": [
    {
      "id": "1647935464530235393",
      "id_data_date": "2023-03-31",
      "id_data_impressions": 0,
      "id_data_app_clicks": 0
    },
    {
      "id": "1647935464530235393",
      "id_data_date": "2023-04-01",
      "id_data_impressions": 327,
      "id_data_app_clicks": 6
    },
    {
      "id": "1647935464530235393",
      "id_data_date": "2023-04-02",
      "id_data_impressions": 55,
      "id_data_app_clicks": 2
    }
  ]
}

Configuration:

"response": {
  "dataRows": {
    "source": "jsonPath",
    "value": "$.data.*"
  },
  "dataAdapters": {
    "items": [
      {
        "type": "time_series",
        "config": {
          "dataPath": "$.data.*",
          "valuesPath": "$.id_data.0.metrics",
          "startTimePath": "$.request.params.start_time",
          "endTimePath": "$.request.params.end_time",
          "granularity": {
            "source": "jsonPath",
            "value": "$.request.params.granularity"
          },
          "dataDestinationPath": "$.id_data",
          "keyName": "date"
        }
      },
      {
        "type": "flatten_values",
        "config": {
          "dataPath": "$.data.*",
          "valuesPath": "$.id_data.*"
        }
      }
    ]
  }
}

Below you can find example field definitions that read from the transformed structure:

[
  {
    "id": "id",
    "label": "ID",
    "dataType": "string.text.value",
    "value": {
      "source": "jsonPath",
      "value": "$.id"
    }
  },
  {
    "id": "date",
    "label": "Date",
    "dataType": "string.time.date",
    "value": {
      "source": "jsonPath",
      "value": "$.id_data_date"
    }
  },
  {
    "id": "impressions",
    "label": "Impressions",
    "dataType": "int.number.value",
    "value": {
      "source": "jsonPath",
      "value": "$.id_data_impressions"
    }
  },
  {
    "id": "app_clicks",
    "label": "App clicks",
    "dataType": "int.number.value",
    "value": {
      "source": "jsonPath",
      "value": "$.id_data_app_clicks"
    }
  }
]