Batched Jobs (Beta)

Batched jobs split a portfolio view or query request into parallel sub-requests and merge the results into a single response. This provides faster computation for large portfolios where a standard synchronous request would time out.

Batched jobs use the same Jobs API endpoint with an additional header.

⚠️

Beta

Batched jobs are in the Addepar Beta Program. Features may be modified or removed at any time.

Overview

📘

Access requirements

Same permissions as the standard Jobs API. The user must have access to the portfolio and view being requested.

How it works

1. Configure a view or query following the guidelines below.
2. Submit the job with the Addepar-Compute-Type: BATCH header.
3. Poll job status via GET /v1/jobs/:id until complete.
4. Download results via GET /v1/jobs/:id/download.

Submit a batched job

POST /v1/jobs

Add the header Addepar-Compute-Type: BATCH to signal batch processing. The request body is identical to a standard job submission.

curl -X POST "https://{firm}.addepar.com/api/v1/jobs" \
  -H "Authorization: Basic {credentials}" \
  -H "Addepar-Firm: 1" \
  -H "Accept: application/vnd.api+json" \
  -H "Content-Type: application/vnd.api+json" \
  -H "Addepar-Compute-Type: BATCH" \
  -d '{
    "data": {
      "type": "jobs",
      "attributes": {
        "job_type": "portfolio_view_results",
        "parameters": {
          "view_id": "74894",
          "portfolio_type": "entity",
          "portfolio_id": "22",
          "output_type": "json",
          "start_date": "2011-12-31",
          "end_date": "2013-01-15"
        }
      }
    }
  }'

{
  "data": {
    "id": "7",
    "type": "jobs",
    "attributes": {
      "job_type": "portfolio_view_results",
      "parameters": {
        "view_id": 74894,
        "portfolio_type": "entity",
        "portfolio_id": "22",
        "output_type": "json",
        "start_date": "2011-12-31",
        "end_date": "2013-01-15"
      },
      "status": "Queued"
    },
    "relationships": {
      "creator": {
        "data": {"type": "users", "id": "44"}
      }
    },
    "links": { "self": "/v1/jobs/7" }
  }
}

Response codes:

• 202 Accepted -- Job created and queued
• 400 Bad Request -- Invalid parameters or unsupported batch configuration
• 403 Forbidden -- Insufficient permissions

Download results

GET /v1/jobs/:id/download

Returns results in the format specified by output_type (JSON, CSV, TSV, or XLSX).

Response codes:

• 200 OK -- Results returned
• 303 See Other -- Follow the Location header to retrieve results
• 404 Not Found -- Job has not completed or produced no results
• 410 Gone -- Results expired (older than 24 hours)

If the view contains unsupported attributes, the job status becomes Error with a description of the incompatibility. Check status via GET /v1/jobs/:id.

Guidelines

Optimize batch performance by minimizing positions per row. Group by position, holding account, or top-level owner rather than broad categories like asset class or country.

Restrictions:

• Summary data applied to a table row is not supported.
• Advanced tables and pivot tables are not supported.

To disable the total row (if unsupported attributes prevent aggregation): uncheck "Show total row" in View Settings, or set disable_total_row = true in the job payload.

Attribute support tiers

Not supported (will cause job failure):

• Benchmark
• Calendar Period Grouping
• Group / Group Type Key (only when >500k groups are enabled)
• Performance Contribution
• Target Allocation / Target Allocation % / Target Allocation Range
• Time Period Grouping
• Transaction-based attributes

Fully supported (non-exhaustive):

Accrued Income, Change In Accrued Income, Creation Date, Display Name, Entity ID, External ID, Group ID, Income/Expenses, Model Type, Net Cash Flow, Net Gain/Loss, Percent of Portfolio, Reference Currency, Total Return, Value.

Partially supported (non-exhaustive):

IRR, TWR.

📘

Related

• Jobs -- Core async job API (status polling, lifecycle)
• Portfolio Views -- View configuration and synchronous results
• Portfolio Query -- Alternative query-based data extraction