Skip to content

Suggestion: run batch jobs in small batches then collate results #7

@antoine-sachet

Description

@antoine-sachet

New feature proposal

Batch requests are limited to 25 documents. Source: https://docs.aws.amazon.com/comprehend/latest/dg/API_BatchDetectSentiment.html

BatchSizeLimitExceededException
The number of documents in the request exceeds the limit of 25. Try your request again with fewer documents.
HTTP Status Code: 400

aws.comprehend handles the exception with a warning and returns an empty list.

My suggestion: when the character vector passed as text has more than 25 elements, the requests should run in batches of 25 and the results should be collated. The behaviour could be enabled by an additional argument batch = TRUE?

I am writing code to that effect anyway. I am happy to share it at some point if we agree this is a good idea.

Metadata

Metadata

Labels

enhancementNew feature or request

Type

No type

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions