[jira] [Commented] (ARROW-14640) reading data from S3

["Nicola Crane (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ARROW-14640?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17442270#comment-17442270 ] 

Nicola Crane commented on ARROW-14640:
--------------------------------------

Hi [~2umxal] , I need to confirm, but I think the issue is that this the proxy settings need to be passed in to the call to {{{}arrow::s3_bucket(){}}}, but this has not yet been implemented.  Your error message there is the same as one which is returned when the user supplies incorrect credentials, but I think it's caused by not having the proxy bits set.  I'll dig into this more and get back to you - thanks for reporting this!

> reading data from S3
> --------------------
>
>                 Key: ARROW-14640
>                 URL: https://issues.apache.org/jira/browse/ARROW-14640
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: R
>    Affects Versions: 6.0.0
>            Reporter: Mikhail Tolmachev
>            Priority: Major
>
> I am trying to read data directly from S3. In my work pipeline I work under proxy, so I set system environments HTTP_PROXY and HTTPS_PROXY. Also I have set AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY.  I have constructed links for reading:
> {code:java}
> s3_uri <- paste0("s3://", accessKeyId, ":", URLencode(secretAccessKey, reserved = T),"@", bucketName, "/", path) 
> s3_uri_short <- paste0("s3://", bucketName, "/", path)
> bucketS3 <- arrow::s3_bucket(bucket = bucketName, access_key = accessKeyId,
>                        secret_key = URLencode(secretAccessKey, reserved = T)){code}
> But none of them worked for arrow::read_parquet
> {code:java}
> df <- arrow::read_parquet(s3_uri)
> df <- arrow::read_parquet(s3_uri_short) 
> df <- arrow::read_parquet(bucketS3$path(x = path)){code}
> Error
> {code:java}
>  IOError: When reading information for key AWS Error [code 15]: No response body. with address : {code}
> How can I fix the problem or maybe get more informative output ? Thanks for attention. 
> I knew about same issue https://issues.apache.org/jira/browse/ARROW-12126 but I can't reopen it.
>  



--
This message was sent by Atlassian Jira
(v8.20.1#820001)