Re: Relationship between PostgreSQL and Datalakes and etc on Azure

From: Steve Midgley <science(at)misuse(dot)org>
To: Shaozhong SHI <shishaozhong(at)gmail(dot)com>
Cc: pgsql-sql <pgsql-sql(at)lists(dot)postgresql(dot)org>
Subject: Re: Relationship between PostgreSQL and Datalakes and etc on Azure
Date: 2021-02-09 16:44:15
Message-ID: CAJexoSJG15aa1AN4ie7Gxk=VSbsW7UazR_atCAxmdKC88n8ndA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-sql

On Tue, Feb 9, 2021 at 1:24 AM Shaozhong SHI <shishaozhong(at)gmail(dot)com> wrote:

> Hi,
> With rapid deployment of Azure datalakes and etc, what is the relationship
> between Azure datalakes and PostgreSQL/PostGIS?
>
> What the advantages has PostgreSQL/PostGIS got over Azure datalakes and
> etc?
>
> Any explanations or articles on this topic?
>
> Regards,
>
> David
>

I'm not clear what tech in Azure tech you're referring to -- there are
various MS SQL solutions out there, and then there are PowerBI type
solutions.

From my perspective, PowerBI offers a lot of high level tooling that makes
data warehouse/analysis type projects really fast/efficient. For analytics
problems, PowerBI can be a very cost effective overall solution that is
worth considering. That said, I don't think you have to give up on Postgres
to get PowerBI working:
https://docs.microsoft.com/en-us/power-query/connectors/postgresql

Regarding the broader "what datalake technology should I use" question --
there will never be a universal right answer, but you are writing to a
Postgres sql language group. The answer for me personally is "Start by
using Postgres, until you can prove that another option will be measurably
better." I've been let down many times by other technologies, and Postgres
just keeps working. I'm actually right now in the process of ripping out a
Sparql/Neptune database solution and replacing it with Postgres. So it
turns out that even for many complex graph queries, I can model the data in
Postgres and get results that are far more performant (2x-4x) than
dedicated graph databases.

So my personal database architecture rule now is: "Default to Postgres, and
optimize with other solutions only when I find something that Postgres
demonstrably can't handle."

Good luck,
Steve

In response to

Browse pgsql-sql by date

  From Date Subject
Next Message Arturo Guadagnin 2021-02-20 14:41:09 embedded composite types
Previous Message Shaozhong SHI 2021-02-09 09:23:32 Relationship between PostgreSQL and Datalakes and etc on Azure