Pagination throws an error against Auora Serverless V1 running MySQL 5.7

[JeffersonSchuler]

At the current time this is more of a question if this should work.

Our situation is we have an AWS Auora v1 running as serverless running MySQL 5.7.

We have a large dataset that exceeds being queried in one call. I have seen in the docs for this library that there is some kind of pagination using server side cursors.

Stepping into the code the method "_start_paginated_query" throws an error here: "self._client.execute_statement(**execute_statement_args)" line 196.

{BadRequestException}An error occurred (BadRequestException) when calling the ExecuteStatement operation: Transaction AQ1a0rmK5xPx2DUvqTwbMHbNhwxGqF5Jz7hBXu732zAwepruvVqGRNLm0DkYoXHF0NX79RCmPvRFySf6NXhXMx9wLihYoIIpswCqo11yQiBoK+UBjxtHWzjMatnv/SYhWxHuZwiLKxwU78KIK7LFCwyEnpO1D9qcfntuyC95WC3+RD/sHww8uMGqQKKHbR9siS1c89xZZ2ga is not found: transaction has been aborted due to an error from a previous call

So I was just wondering if this is even supposed to work with this type of server configuration.

Thanks, I really like this library.

[kislyuk]

Thanks. Yep, unfortunately this is a known issue. Please see the comment here: cloud-utils/sqlalchemy-aurora-data-api#38 (comment)

I am in a holding pattern with AWS support on this, but chances are I will have to remove the automatic backoff logic altogether since AWS has decided that any active transaction must be aborted on any error - and there is no way to connect multiple statements into a single database connection context without referencing a transaction. You can still use pagination manually with LIMIT..OFFSET or even tap into the internal API of this package (__start_paginated_query()) but the automatic pagination no longer works.