Dockerfile in the Custom Docker File Approach for R Example Needs to be Updated

## The Dockerfiles in the following examples are not updated.

example:
- https://github.com/Azure/azureml-examples/tree/main/cli/jobs/single-step/r/iris
- https://github.com/Azure/azureml-examples/tree/main/cli/jobs/single-step/r/accidents

description:
The job fails when creating it using the existing sample. It is showing the following errors in the logs.

AttributeError module 'dateutil' has no attribute 'parser'
System command 'mlflow' failed, exit status 1, stdout & stderr were printedWarning messages

After several tests, I found the following 2 issues.
- The base image is still referencing Ubuntu version 18.04
- mlflow only compatible with Python 3.7

As a workaround, I updated the Dockerfile and used the latest base image [Latest Base Image](https://hub.docker.com/r/rocker/tidyverse/tags) and add the pip install mlflow so that the image gets the latest mlflow version.

## Additional context

-

Start with the Dockerfile files in cli/jobs/single-step/r/iris and cli/jobs/single-step/r/accidents, then run the sample job creation flow to reproduce the logged dateutil and mlflow failures. Compare the current base image and installed dependencies with the reported workaround; done means both R examples create and run successfully without those errors.

Written by the indexing model from the issue text.