Getting started with Azure Synapse Analytics

[@travissondgerath4543]

More, great overview but it would be great to really dive into the capabilities of this platform.

[@eziola]

Thanks for starting at the clicks and basics. Something I'm interested in, is say I've got monthly .csv drops and want to build refreshable Power BI visualizations using that data. Should I just drop the .csv files in the data lake each month and loop through them in Power BI? Or could I set it up so that the .csv files are imported into a SQL table that Power BI connects to? Thoughts? Possible video idea? I think this is a common design question for Self Service BI developers.

[@nickberesford]

Working in a large organisation, lots of that stuff (Resource Group etc.) has already been set up, and we already have data in our ADLS Gen 2. Where I'm struggling is coming at Synapse from an analysis point of view (Excel, and more recently Power BI), there just seems to be so many different concepts to grasp at once, with similar terminology, that it gets confusing very quickly. All I want to do (for now) is connect to data in the data lake, turn it into a really useful datast for self-service analysis by the business, and then serve that up in Power BI. I wonder how many Synapse 'beginners' are actually trying to set up Synapse starting from a blank canvas, and how many have been handed a new platform that has been set up by 'IT', and are expected to get on with delivering advanced analytics despite it feeling like an alien environment!

[@ncbshiva]

This is great, can you show us how to create dedicated sql pool and what is the difference between serverless and dedicated sql pools in Azure Synapse also which sql pool should we use with some small examples ?

[@hugobohorquez7450]

excellent video. Short, sweet and informative. Thanks!

[@spilledgraphics]

It´s all about learning the clicks. Amen.

[@culpritdesign]

Synapse is really coming together. Thanks for sharing!

[@chinmaygarg281]

Will this incur any additional charge on my Azure subscription?

[@Chiz_1]

Say you come from Power BI, small-mid sized business. You have a couple of sources, operational DB and some excel files. You have previously created power bi dataflows for the most used tables and then models on top of that, and are now interested to move it all to Synapse. How would you go about it?

[@saratchandra7388]

You have dataflows in azure synapse may we can leverage that one.then pbi is present .but that would increase cost.better to use current model now

[@zxccxz164]

get ready to spend WAY over what you budget on this project. I have seen it happen already

[@matthewm9595]

Great video, this ASA looks like it tries to bring more things into one place. Can it replace databricks and application insights for example? Why would I use this and not other resources? Thanks!

[@harshithasureshvanajakshi6976]

I am trying to set up a new workspace for synapse analytics and when I did I get an unreachable status on Built-in SQL pool, if you could let me know why or what I could do to fix it, that would be great please

[@wookieeMan06]

Question: If you create a database or a data-lake store here. Is it a normal Azure data lake component (need a subscription before)? or it part of the service?

[@dataworld9783]

is there any specific certification available for Synapse?

[@josuelabaut9969]

Amazing video Patrick! I would like to see some transformation using notebooks with PySpark can I use that with my existing Synapse?

[@michaeld9682]

Thank you. Great job

[@RonaldPostelmans]

Hi Adam, nice video. Have you any idea from own experience if you can create large star models that still are fast with Synapse serverless. I think it is amazing, but i'v heard from other people that if you have large Fact tables for example invoicing table with 40 million recoreds, synapse serverless lacks performance. If thats the case en you can only use it for dimension tables who most of the time hold much less records. then you have to have 2 sources, for example 1 Synapse serverless for dimension tables and 1 azure sql database for Fact tables for performance reasons. Please, I would like to hear what your experience is.

[@zxccxz164]

My experience, less than a month....sadly is performance is unreliable in that one week queries were 7 times faster than on prem. Than the following week 10 SLOWER than on prem for no apparent reason. I would like a clear explanation of why SYNAPSE should be used over SSAS (on prem or Azure)

[@alexfernandodossantossilva4785]

Man, why don't you compare Data Factory and Synapse? Which tool should use to orchestrate?

[@radekmiszczyk9401]

Thank you Guys for what you are doing. It's great that besides the PowerBI, you took also Azure Synapse on the plate. I'm sure it will be on the same, high level :-) it would be great to show how to transform data using Spark and their notebooks.

[@stephandemedts8894]

I really like this solution but I don't really know how to kind of sell it to my boss so that we move from old MS SQL warehouse to an Azure one. I think its PBI connectivity seems really useful and a lot of additional features that previous solutions were not providing.

[@zxccxz164]

what solution did he show?

[@preethamema7973]

Lets say a company has many application systems and if each system need to maintain only 2 yrs data in their own application database and rest of 10+ yrs data to ADL. Can the users of system 1 access archived data of 10+ yrs using PBI, users of system 2 access archived data of 10+ yrs using Qlik and at the same time the data scientists can access archived data from system 1, system 2, system 3 which are all saved in ADL?

[@zxccxz164]

Can they SHARE the tabular model?

[@preethamema7973]

@@zxccxz164 The reason for migrating to ADL is for cost of storage in D365 for all 12+ yrs is high and there are existing reports which have been created using PBI and Qlik on relational DBs. Tabular model will be our end state across enterprise which will be achieved progressively in less than 2 yrs. For now, keeping in view of the existing reports and D365 cost consideration we are thinking to move to ADL as is

[@preethamema7973]

@@zxccxz164 Operational users should only see their data, management users can see shared data. Data scientists should have access to raw and processed data

[@chilliyou]

Biggest challenge is investing and subscribing. Need clarity how billing works and how you control it.

[@zxccxz164]

hahahhaahhahaha......you mean hook you in.....and inflate charges each year or half year as cost are exploding everywhere? yeah, i would think people would want to discuss this.

[@miragliag]

Is there a low-cost version or per user basis version available?

[@keen8five]

most Synapse functionality is usage-based without fixed cost

[@saisatya5525]

you are always awesome..

[@keen8five]

I disagree Patrick. If you want to do meaningful stuff with the platform just remembering clicks will not get you far. Synapse is very powerful but also complicated to comprehend - especially how to leverage all its functionality properly.

[@zxccxz164]

finally someone not paid by MS to tout their inflationary sub model

[@evnpresson7258]

Wow fastest talking ever haha

[@zxccxz164]

STOP please. Tell me WHY we need this (without the reason being, because MS is on HUGE push to make everything subscription based where they can alter the prices at whim)

[@ilia_playzz]

So many producer at the end of the video but it was absolutely useless