Hi @Rajesh-Singh, we didn't do a formal comparison yet. In general,
using Kusto you can store the embeddings in your table in your own db,
and you have full control of it (w.r.t. retention, cross region and
more). We are also adding (IVFFlat) indexing of the vectors to further
accelerate the search ...
@adieldar , great article. I'm tempted to do a POC on ADX based RAG
pattern. Do you have comparison of these results with Azure AI Search as
well? it would be interesting to see how these two products can
complement each other, and which use cases these products fits into.
Delta has become a standard format. But as an Azure user, I am little
confused about the many options I have to convert my ADX data in delta
format now. I could use ADX external tables to export to delta using
continous export (> my favourite)use Data Factory pipelines which
require a databricks ins...
I am a little bit disappointed of these base queries. I expected that a
base query is executed once and the results can be re-used several
times. So the base query would simply pass the result data to multiple
tiles. So I created a very complex query, defined a base query and
created 30 tiles referr...
Thanks, Dany. Good read! Did you change the partition policy or is your
timestamp column (CreatedAt) identical to the ingestion date (which is
the default partition column)?
Latest Comments