Database Reference
In-Depth Information
Figure 10.15
Cleaning up directories in Hadoop with -rmr
PerhapsamorecompleteextensiontoCETASwouldbe
WITH (TRUNCATE,
DROP_EXISTING)
. This would both remove the existing external table and
fire the Hadoop file system
-rmr
operation to clean up the data residing in
Hadoop before the next export. However, at this stage neither
TRUNCATE
nor
DROP_EXISTING
exist at this moment.
Business Use Cases for Polybase Today
One of the comments I've heard several times is that Microsoft tends to only
talk about what is coming down the line. Although this is both interesting
and brings much excitement, it often also leads to frustration as readers,
listeners and viewers often end up feeling that there is no immediate
use-case for the technology. I am going to try to address that in this section,
at least in part. Therefore, in the following subsections, I discuss some of the
ideas I've had for how to use Polybase, based purely on the building blocks
we have today.
Archiving and Audit
The first scenario I want to suggest is an extension to the archiving solution
Thomas Kejser proposed and that I highlighted earlier in this chapter.
Thomas suggested that Hadoop and Hive could be a simple place to store
source system data, which could both be easily queried and with Polybase
now very easily rehydrated if needed. This would enable all sorts of
warehouse replay functionality.
