NEWS.md
httr to httr2
sso credentials by calling aws cli (#844)debug and trace. This is to prevent info level being saturated by api calls.PawsStreamHandler, allows paws to handle aws stream event (#842). Thankyou to @hadley for developing the initial solution in httr2.s3_unmarshal_select_object_content in favour or new streamhandlerparse_url, parse_query_string and build_url to cpp for performance improvement.url is set to lower case before signatureaws-sdk-js to aws-sdk-python boto3.handle_copy_source_param encoding redirected calls.handle_copy_source_param removing attributesunix_time ensure seconds is numeric (#804), thanks to @joseale2310 and @lyschoening for raising issue.x-amz-* headers (#815) thanks to @cgostic for raising issueCopySource in operation CopyObject (#819)CopySource for CopyObject operations (#819). This is to align with boto3 implementation.
library(paws)
client <- s3()
bucket = "BUCKET"
key = "%01file%/output.txt"
resp <- client$put_object(
Bucket = bucket,
Key = key,
Body = charToRaw("helloworld")
)
client$copy_object(
Bucket = bucket,
Key = "file_out_1.txt",
CopySource = sprintf("/%s/%s", bucket, key)
)
client$copy_object(
Bucket = bucket,
Key = "file_out_2.txt",
CopySource = list(
Bucket = bucket,
Key = key
)
)uuid to cpp for performance improvement.transpose to correctly parse lists with empty first elements (#791), thanks to @FMKerckhof for raising issue.sso (#793)
# previous error message format
svc <- paws.storage::s3()
response <- svc$get_object(
Bucket = "<bucket>",
Key = "<key>",
IfNoneMatch = "<etag>"
)
#> Error: SerializationError (HTTP 304). failed to read from query HTTP response body
# new error message format
client <- paws.storage::s3()
resp <- client$get_object(
Bucket = "<bucket>",
Key = "<key>",
IfNoneMatch = "<etag>"
)
#> Error: SerializationError (HTTP 304). An error occurred when calling the GetObject operation: Not Modifiedxml_parse to correctly parse empty elements (#783) thanks to @stevepowell99 for raising issuerestxml_unmarshal by x3rest_unmarshal_location_elements only skip header if location is not found (#761)endpoint_url in config file and environmental variables (#764), thanks to @James-G-Hill for raising issuejson_convert_string to cpp for significant performance improvement (#754)resolve_endpoint to cpp to get a x2 performance enhancementspaginate.jmespath_index
aws-global region when resolving endpoint (#730). Thanks to @atheriel for identifying the issue.AWS_CONTAINER_CREDENTIALS_FULL_URI environmental variable. This supports to services like sagemaker serverless endpoints (#737). Thanks to @ncullen93 for raising issue and testing.AWS_* (#706, @fh-mthomson)read_ini
read_ini for improved performanceweb_identity_token_file in AWS config file thanks to @liuquinlin for implementation.NULL method as.POSIXct.default for older versions of R (#698). Thanks to @gp-dev-pw for identifying issueescape function by removing protocol check (#701)generate_presigned_url (#572)xml_build in previous version (#569).AssumeRoleWithWebIdentity (#477, thanks to @fh-mthomson for contribution)AWS_CREDENTIAL_EXPIRATION is optional when using AWS_SESSION_TOKEN (#549)build_query_string by exiting safely when parameter is empty.cpp, and dependency Rcpp.xml_build to allow empty characters from operations.escape with new helper function paws_url_encoder (1000x improvement).paws_url_decoder (780x improvement).aws sso login --profile my-profile.LastModified from get_object.issue network request function to avoid CRAN check errors.AWS_DEFAULT_REGION in addition to AWS_REGION.do.call, including any user-supplied configuration to the service, e.g. a custom endpoint.s3_put_bucket_lifecycle_configuration. Fixes #438.s3 <- paws::s3(config = list(close_connection = TRUE)). Fixes #431.Use path style URLs for S3 when using custom endpoints, (e.g. localhost:9000/mybucket). This is useful when using software like MinIO. When not using custom endpoints, Paws will in general use “virtual hosted” URLs (e.g. mybucket.s3.amazonaws.com) except in special cases.
Use path style URLs for S3 when using the new s3_force_path_style option when calling paws::s3(), e.g.:
s3 <- paws::s3(config = list(
s3_force_path_style = TRUE
))Add support and a default value for timeout, how long to wait in seconds for an acknowledgement of an HTTP request before failing. Default = 60. This is user configurable by adding arguments to the service call, e.g.
svc <- paws::svc(config = list(timeout = 10))Don’t skip lower case field names for REST location elements. For example, lexmodelbuildingservice$create_bot_version(name, checksum) previously failed due incorrectly skipping the name field.
mfa_serial shared configuration file item.AWS_CONFIG_FILE, AWS_SHARED_CREDENTIALS_FILE, AWS_CREDENTIAL_EXPIRATION, and AWS_EC2_METADATA_DISABLED environment variables. See https://github.com/paws-r/paws/blob/main/docs/credentials.md for explanations of all settings.s3$put_object will now read in files when given file paths for the Body parameter. Previously it would accept only blobs, e.g. from readBin.paws_error in addition to their other classes, e.g. http_400, error, condition.role_arn, credential_source, and source_profile.credential_process.numeric objects rather than 32-bit integer objects.s3$select_object_content. This operation allows you to query a data file in S3 using SQL."Me voy a casa mañana.".paws::svc(config = list(credentials = list(profile = "my-profile"))). Previously the profile was not used to get credentials.svc <- paws::svc(config = list(region = "us-west-1")).new_service, namely each endpoint should look like list(endpoint = "foo", global = FALSE).