Independent leaderboard for analytics agents @ https://teapawt.net/t.co/UOv6GoVeFM. Also building semglot: like sqlglot, but for semantic layers. Open source.
Ontario, Canada
Joined February 2013
- Tweets44
- Following21
- Followers13
- Likes5
Semglot: sqlglot for semantic layers. We built a small open-source tool and I think a lot of you will want it.
semglot takes one semantic model and transpiles it into every format your tools expect: dbt, Snowflake Cortex, semantic views, nao.
github.com/benchouse/semglot
1
1
It's v0.1.0 and rough in places. Go, MIT, one dependency. dbt is the source today, six targets and growing, and every mapping is documented so new formats are easy to add.
1
1
Repo: github.com/benchouse/semglot The leaderboard it came out of: benchouse.ai
If you maintain a semantic model in more than one tool, kick the tires and tell me what breaks.
1
An agent now answers a data question in seconds for under $2. I think it's heading under 10¢. An analyst cost a lot more.
But cheap only matters if it's right. What counts is cost per correct answer, not cost per question.
benchouse.ai, the analytics agent benchmark.
7
#SQL Tip #4
3NF, third normal form is a great way to model transactional databases! No duplication of data (except PKs) in the minimum amount of tables.
If you don't know about it: youtube.com/watch?v=aAx_JoED…
#dataengineering #databases #DataScientists #datamodeling
Data Engineers Must Take a Customer-First Mindset, interesting and short read that is spot on, on how the role of a data engineer is evolving:
thenewstack.io/data-engineer…
#DataEngineering
#DataScience
#SQL Tip #3
If you think you have to use an outer join -> 95% of the time you don't.
Only when you know you need to -> 50% of time you do.
Avoid them until you cannot avoid them, as they make your code less readable, less understandable & buggy.
#DataEngineering #DataScience
André Baaij retweeted
In fact, #DataEngineering opportunity for our target customers is much bigger than data science. This pyramid not only illustrates the importance of the foundational work but also the volume. So, we decided to make our #ETL service more independent as
👉 @handoffcloud
#SQL Tip #2, store your true/false as bool values and name them correctly like: has_started, is_active.
Now you can query all active training's that have not started yet like this:
SELECT *
FROM TRAININGS
WHERE IS_ACTIVE
AND NOT HAS_STARTED
false & NULL are handled similar
André Baaij retweeted
Software Engineer Vs. Data Engineer - Is There A Difference? #datascience #dataengineering #sql #analytics #python #java #coding youtu.be/0wjngjhmtJA via @YouTube
Definitely going to check out this in more detail. Datavault and dbt combined in a package seems like it could be a win-win!
This Post is from an account that no longer exists. Learn more
Scrolling through #DataEngineering where half of the tweets are job postings, can you all post some content please 😂
#SQL Tip #1: Be sure to use transactions when testing update/delete code in prod.
Great tutorial on how to use it right here: sqlitetutorial.net/sqlite-tr…
PS. If your database doesn't support transactions, you might want to rethink your database choice :)
#DataEngineering #Database
Which IDE do you use to work with your
#DataEngineering projects?
My go to favourite is datagrip from @jetbrains, it makes me at least 20% more effective. #DataOps #DataScience
2
André Baaij retweeted
I was gonna keep quiet about it, but HERE WE GO
I got a new entry level job 🥳🥳🥳
Job Title: Data Engineer
Company: DAT solutions
Pay: 90K
Benefits: Yes, Regular stuff
Location: Remote for now
I started yesterday. I kinda got lucky with this job 😀.
How I got the job 👇👇👇
#snowflake tip: if you are wondering why your queries are not finishing, the warehouse queue might be a bit full 🤨
As far as I could see this is the only place you can find this info...
#dataengineering #Database #performance #DBA
Great article on which tool to use to create analytical dashboards: medium.datadriveninvestor.co…
TLDR: #streamlit for small companies and #dash for enterprises #panel might be the underdog! #DataScience #DataEngineering #Dashboard #Analytics
1
