---json
{
    "page_id": "rdsmd9ruqzgoaxl9wefva"
}
---
====== Data Operation - Define ======


===== Define =====
''define'' is a [[:docs:flow:supplier|supplier]] [[op|operation]] that defines one or more inline [[docs:resource:resource|data resources]]
in a [[docs:flow:pipeline|pipeline file]].


Note that a [[select|select operation]] may not return any [[docs:resource:resource|data resources]] but the ''define'' operation must return always the data resource that it defines.


This is a first step of a [[:docs:flow:pipeline|pipeline]] (known as the [[docs:flow:supplier|supplier step]]) because it supplies [[docs:resource:resource|data resources]] to the [[:docs:flow:intermediate|others operations (known as intermediate)]] in the pipeline.

===== Example =====

  * [[howto:pipeline:define]]

===== Arguments =====
The ''define'' operation accepts the following [[:docs:flow:argument|arguments]].

^ Arguments ^ Default ^ Definition ^
| ''data-resource'' | | One [[#data resource definition]] |
| ''data-resources'' | | A list of [[#data resource definition]]   |
| ''processing-type'' | ''batch'' | The [[:docs:flow:processing-type|processing type]] (''batch'' or ''stream'')  |


==== Data resource Definition ====

The following argument defines one data resource and can be set:
  * once in the `data-resource` argument
  * multiple time in the `data-resources` argument

You can define a data resource by:
  * `data-uri`
  * or by `data` (inline rows)
but not both

^ Arguments ^ Mandatory ^ Definition ^
^ One resource definition by Data Uri ^^^
| ''data-uri'' | only if ''data'' is not defined | The [[docs:resource:data_uri|data uri]] is the location of the resource |
| ''media-type'' |  | The [[docs:resource:media-type|media type]] of the data uri resource |
^ One resource definition by Data ^^^
| ''data'' | only if ''data-uri'' is not defined | The ''data'' as a list of [[docs:resource:record|records]] (the data are in the inline) |
^ The optional data attributes ^^^
| ''data-def'' | no | The data attributes in a [[docs:resource:data-definition|data definition format]] |


==== Processing Type ====

  * In a ''batch'' [[:docs:flow:processing-type|processing type]], the data resources are sent in the pipeline only once.
  * In a ''stream'' [[:docs:flow:stream_pipeline|stream processing type]], the same data resources are rebuild at poll time and resend.