Copy page

Transformations

Pivoting and Reshaping

Use these to rotate or restructure your data when the shape itself needs to change.

---

Pivot — DF_PIVOT

Rotate your data so that unique values from one column become the column headers.

Use this when: you have a table with a "category" column and you want each category to become its own column — turning rows into columns.

Before

Code
[
  { "region": "North", "metric": "revenue", "value": 500 },
  { "region": "North", "metric": "cost",    "value": 300 },
  { "region": "South", "metric": "revenue", "value": 420 },
  { "region": "South", "metric": "cost",    "value": 250 }
]

After — pivoted on metric

Code
[
  { "region": "North", "cost": 300, "revenue": 500 },
  { "region": "South", "cost": 250, "revenue": 420 }
]

Code
{
  "type": "DF_PIVOT",
  "config": {
    "rowField": "region",
    "columnField": "metric",
    "valueField": "value"
  }
}

---

Rows to Fields — DF_ROWS_TO_FIELDS

Convert key-value rows into columns — each unique key becomes a column name and its value populates the cell.

Use this when: your data comes back as a list of { "key": "...", "value": "..." } pairs but you want a flat row with named columns.

Before

Code
[
  { "key": "temperature", "value": 24 },
  { "key": "humidity",    "value": 60 },
  { "key": "co2",         "value": 412 }
]

After

Code
{ "temperature": 24, "humidity": 60, "co2": 412 }

Code
{
  "type": "DF_ROWS_TO_FIELDS",
  "config": {
    "nameField": "key",
    "valueField": "value"
  }
}

---

Labels to Fields — DF_LABELS_TO_FIELDS

Promote a column's label metadata into its own columns (or rows).

Use this when: your series carry labels — like host or region tags attached to a metric — and you want those labels to appear as ordinary columns you can group, filter, or display.

Before — a value series labelled with host and region

Code
[
  { "value": 24, "host": "node-1", "region": "sg" },
  { "value": 31, "host": "node-2", "region": "us" }
]

After — labels lifted into columns

Code
[
  { "host": "node-1", "region": "sg", "value": 24 },
  { "host": "node-2", "region": "us", "value": 31 }
]

Code
{
  "type": "DF_LABELS_TO_FIELDS",
  "config": {
    "labelKeys": ["host", "region"],
    "mode": "columns"
  }
}

Omit labelKeys to promote every label. Set "mode": "rows" to lay the labels out as rows instead of columns.

---

Extract Labels — DF_EXTRACT_LABELS

Split a structured metric name into separate label columns.

Use this when: your metric names pack several pieces into one string — Prometheus-style (http_requests{host="node-1"}), Graphite-style (servers.node1.cpu.load), or a custom pattern — and you want each piece in its own column.

Before (Graphite-style names)

Code
[
  { "metric": "servers.node1.cpu", "value": 0.62 },
  { "metric": "servers.node2.cpu", "value": 0.41 }
]

After — name segments split out

Code
[
  { "group": "servers", "host": "node1", "resource": "cpu", "value": 0.62 },
  { "group": "servers", "host": "node2", "resource": "cpu", "value": 0.41 }
]

Code
{
  "type": "DF_EXTRACT_LABELS",
  "config": {
    "metricField": "metric",
    "pattern": "graphite",
    "separator": ".",
    "segmentNames": ["group", "host", "resource"],
    "keepOriginal": false
  }
}

Set pattern to "prometheus", "graphite", or "custom". For "custom", supply a named-group customRegex. Set "keepOriginal": true to keep the original metric column alongside the extracted ones.