Arguments | Category | Description | Default Values | Permitted Values | Data Type | Charset (Applicable for String) | Case Sensitive |
---|---|---|---|---|---|---|---|
API_TYPE | Optional | Specifies the API to be used. When working with Google Vertex AI, the default and only permitted value is '"vertex-ai'.
|
vertex-ai | vertex-ai | String | Unicode | No |
AUTHORIZATION | Required | Specifies a JSON string representing key or token to access endpoint. | None | JSON String of the following format: '{"key" : "gcp-vertex-access-token", "region" : "gcp region string", "project" : "google project name"}' region and project are optional if a complete URL is specified in the ENDPOINT argument.
|
String | Unicode | No, JSON keys are not case sensitive. |
ENDPOINT | Required | Specifies the name of Google Vertex AI endpoint, or the scoring URL. | None |
|
String | Unicode | Yes, value is used as is for connecting to Google Vertex AI. |
CONTENT_TYPE | Optional | Specifies content type for sending input data to Google Vertex AI. | json |
|
VARCHAR | Unicode | No, argument is not case sensitive. |
CONTENT_FORMAT | Optional | Shows how the rows should be formatted. This element is used only when CONTENT_TYPE is json.
For example: "{ "data" : [ [ "%row" ] ] }" This will make API_Request to format row as follows: (assuming 10 columns):
{"data": [ [1,2,3,4,5,6,7,8,9,10], [10,9,8,7,6,5,4,3,2,1] ]} |
Empty string | Value should be json string with %row placeholder where Teradata row will be placed. "%row" must be an array and it's parent may be array if model supports multiple rows. | VARCHAR | Unicode | Yes, value will be used as is for creating row. |
KEY_START_INDEX | Optional | Specifies the column index for the key in the input column. Column index count starts from '0'. For example: For SELECT col1,col2, key1, key2 from table, you should use '2' so that In-database Function will consider key1, key2 to form key to be sent to Azure. |
NA | NA | VARCHAR | Unicode | No |