Create Jobs from a CSV file

Uploads a CSV file and creates Jobs from its records. The response contains created Jobs and validation errors.

Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…

CSV layout

Upload a comma-separated file as the file field of a multipart form request. Each record after the skipped records creates a Job. Columns are selected by zero-based positions, not header names. The defaults are query_index=0, from_date_index=1, and to_date_index=2. Set properties_index to include a column containing JSON Job properties; it has no default. Negative indexes are ignored.

skip_records defaults to 1, so the first record is skipped by default. Set it to 0 for a file without a header. The query must be nonempty. Date cells can be empty or null; otherwise they must contain a date the server can parse, such as an ISO 8601 date. Dates without a time zone are assumed to be UTC. When both dates are present, the start must precede the end. Quote CSV cells containing commas or double quotes, and escape embedded double quotes by doubling them.

For example, this file places the query in column 0 and the collection dates in columns 1 and 2:

query,from_date,to_date
electric vehicles,2026-06-01T00:00:00Z,2026-06-02T00:00:00Z

For this layout, set query_index=0, from_date_index=1, and to_date_index=2. Set skip_records=1 to skip the header. The column names in this example are labels; the indexes determine which columns the endpoint reads.

The job_type query parameter applies to all Jobs created from the file. The backend recognizes these Job types:

ValueBehavior
oneTimeRun once.
periodicRun periodically.
continuousRun continuously.
mergedPeriodicRun periodically, merging missed periods.
dynamicPeriodicRun periodically, adjusting the period based on returned data.
queuedPeriodicRun periodically, one Job at a time.
asyncPeriodicRun periodically, with parallel backfill.

Type matching ignores case and underscores. The validator requires a valid schedule for periodic and mergedPeriodic. The legacy runNow value maps to oneTime. An unrecognized type maps to oneTime; supply an explicit type to avoid an unintended result. The file handler sets priority to normal.

The response can contain both created Jobs and validation errors: valid rows can be created even when other rows fail validation. Check total_created_jobs and errors before retrying a file. Malformed CSV, dates, or JSON properties can fail the request during parsing before Jobs are created.

Path Params
string
required

The pipeline ID.

string
required

The component step ID, returned as step_id in Job and work item responses.

Query Params
string
required

Data source. Required.

int32

Zero-based query column index.

int32

Zero-based start date column index.

int32

Zero-based end date column index.

int32

Zero-based JSON job properties column index.

string

Job type: oneTime, periodic, continuous, mergedPeriodic, dynamicPeriodic, queuedPeriodic, or asyncPeriodic. Matching ignores case and underscores. Unrecognized values map to oneTime; legacy runNow also maps to oneTime.

string

Schedule for periodic Jobs.

string

Prefix used for generated Job names.

int32
Defaults to 1

Number of CSV records to skip.

boolean
Defaults to true

Create Jobs ready to run.

Body Params
file
required
Headers
string
enum
Defaults to application/json

Generated from available response content types

Allowed:
Responses

401

Unauthorized.

Language
Credentials
Header
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json
text/plain