Postprocess is the power tool. It’s a JavaScript block that runs after all queries and computed fields, giving you access to the entire result set at once. Need to group, sort, pivot, aggregate, or reshape your data? This is where it happens.

Used by 16% of HRM components — not as common as computed fields, but essential for the complex ones.

📋 Syntax

postprocess: |-
  // JavaScript code
  // $result is the full array of data rows
  // Must reassign $result at the end
  $result = $result.map(row => ({
    ...row,
    newField: row.existingField * 2
  }));

⚠️ Critical Rules

  1. $result must always be an Array at the end — never an Object
  2. $result is mutable — you can reassign it
  3. All computed fields are available on each row in $result
  4. Postprocess runs once per query context, not per row

Example: Population Change Calculation

This postprocess groups data by name, calculates percentage changes across years, and creates delta labels:

postprocess: |-
  const createDeltaLabel = (difference) => {
    if (Math.abs(difference) < 0.05) {
      return `<span style="color: lightgrey;">⬭</span> ${Math.abs(difference).toFixed(1).replace('.', ',')} %`;
    } else if (difference < 0) {
      return `<span style="color: red;">▼</span> ${Math.abs(difference).toFixed(1).replace('.', ',')} %`;
    } else {
      return `<span style="color: blue;">▲</span> ${Math.abs(difference).toFixed(1).replace('.', ',')} %`;
    }
  };

  const calcChange = (data, yearStart, yearEnd) => {
    const popStart = data.find(d =>
      d['cube.stichtag'].startsWith(yearStart)
    )?.['cube.population619'] || 0;
    const popEnd = data.find(d =>
      d['cube.stichtag'].startsWith(yearEnd)
    )?.['cube.population619'] || 0;
    return popStart === 0 ? 0 : ((popEnd - popStart) / popStart) * 100;
  };

  // Group data by name
  const groupedData = {};
  $result.forEach(row => {
    if (!groupedData[row.name]) {
      groupedData[row.name] = [];
    }
    groupedData[row.name].push(row);
  });

  // Aggregate per group
  $result = Object.entries(groupedData).map(([name, data]) => {
    const change2000_2023 = calcChange(data, '2000', '2023');
    const change2020_2023 = calcChange(data, '2020', '2023');
    return {
      'cube.populationChange': change2000_2023,
      'cube.populationChange_2020': Math.abs(change2020_2023),
      'cube.delta_label': createDeltaLabel(change2020_2023),
      'name': name
    };
  });

🧩 Common Patterns

Flatten Nested Data

When API responses contain nested objects (e.g., TIE salary_quantiles):

postprocess: |-
  $result = $result.map(row => ({
    ...row,
    q10: row.salary_quantiles?.q10 || 0,
    q50: row.salary_quantiles?.q50 || 0,
    q90: row.salary_quantiles?.q90 || 0
  }));

Top-N Filtering

postprocess: |-
  $result = $result
    .sort((a, b) => b['cube.weightedShare'] - a['cube.weightedShare'])
    .slice(0, 10);

Data Pivoting

Transform rows into a different structure:

postprocess: |-
  const grouped = {};
  $result.forEach(row => {
    const key = row['cube.category'];
    if (!grouped[key]) grouped[key] = { name: key, values: [] };
    grouped[key].values.push(row['cube.value']);
  });
  $result = Object.values(grouped);

German Number Formatting

postprocess: |-
  $result = $result.map(row => ({
    ...row,
    formattedValue: row['cube.value'].toFixed(1).replace('.', ',') + ' %'
  }));

⚡ vs. Computed Fields — When to use which?

Feature Computed Fields Postprocess
Runs Per row Once on full result
Access Single row + arguments All rows ($result)
Can filter/sort No Yes
Can aggregate No Yes
Can change row count No Yes
Execution order Before postprocess After computed fields

Rule of thumb: Use computed fields for per-row transformations, postprocess for aggregations and reshaping.