Extract, transform, load
177
MENTIONS
88
EPISODES
72
PODCASTS
Search complete. 177 mentions across 88 episodes found for "Extract, transform, load".
Sep 12, 2026
Luis Salazar BetancourtGUEST
We can see here, for instance, in the middle, what kind of data we structure for instance, or some other more sophisticated metrics and graphics, so that we can get it finalized to the machine learning input that an AI can understand, a machine learning can understand.
Luis Salazar BetancourtGUEST
And then when you have this, you want to focus now on when we implement the extraction and transformation of what we call the ETL pipeline, we want to play the same for the CAD file.
Luis Salazar BetancourtGUEST
So we transform the CAD into a machine learning input so that then you have all those libraries that are very powerful to use and to actually learn something.
Luis Salazar BetancourtGUEST
So the transformation in the middle, we just take care of the parallel dictation.
Jeremy Von HalleGUEST
We've got a few other tools that we can connect to as well.
Jeremy Von HalleGUEST
But I think for us, we've been able to create the right ETL and reverse ETL layer into Salesforce where we can get a lot of the summary information.
Jeremy Von HalleGUEST
Obviously, we're not using Salesforce as a data warehouse, but we can get some of the key information that would help drive things like customer health or adoption or usage into that.
Jeremy Von HalleGUEST
So it becomes a lot more simple for us to run things on top of it.
Pedro HolandaGUEST
Maybe Guillermo knows.
Guillermo Sanchez-DionisGUEST
There is a Spark writer, I think, specifically that was developed by MotherDuck because some of their customers were using Spark for ETL.
Guillermo Sanchez-DionisGUEST
So they thought that it would make sense to have a Spark writer into Duck Lake.
Guillermo Sanchez-DionisGUEST
I see.
Kumuda SreenivasaGUEST
They take off the screws, they do everything, small, small parts, everything.
Kumuda SreenivasaGUEST
So We have enterprises, database, cloud platforms, analytic tools.
Kumuda SreenivasaGUEST
So we use Airflow as orchestration for ETLs and ELT jobs.
Kumuda SreenivasaGUEST
And we do data quality checks and production reporting and file processing.
Kumuda SreenivasaGUEST
We operate almost on thousands of DAGs and run across different schedules and business processes.
Ran ChenHOST
Be alert for trick questions suggesting you fix data types in Desktop Data View or DAX calculated columns.
Ran ChenHOST
The PL300 best practice is to always perform type conversions in Power Query as early as possible in the ETL pipeline, to optimize query folding, reduce data engine load, and prevent refresh errors.
Ran ChenHOST
Also, do not confuse simple type conversion with locale-aware conversion when foreign regional formats are involved.
Ran ChenHOST
For free practice questions, AI-powered explanations, and more exam prep tools, visit OpenExamPrep.com. That's OpenExamPrep, all one word, dot com.
Jonathan PalmerGUEST
There was shared ownership.
Jonathan PalmerGUEST
You don't just think about charts and you just think about ETL.
Jonathan PalmerGUEST
No, you both need to think about both these things.
Jonathan PalmerGUEST
You may have specialization.
Dumky de WildeHOST
But when I was working a lot with, uh, with DBT for, um, for clients as a, as a consultant, um, you would have data in your data lake, you would data, have data in a Postgres database, and DuckDB would be a very good way to just combine all of these data sources.
Dumky de WildeHOST
It's like you don't even need some kind of ETL tool if-
Mehdi OuazzaHOST
Yeah
Dumky de WildeHOST
... you can connect to a lot of these sources directly with, uh, with DuckDB.
Ash HuntHOST
So I do think, I completely agree, the legacy approaches to detection are already falling by the wayside.
Ash HuntHOST
And you can see that even in approaches to data, right? Instead of large ETL projects or shifting it into SIEM or even a UDP, a unified data platform, we're seeing approaches now where we're being able to read and interpret data in situ wherever it exists without having to move it and building controls based off of that centralized analytical view and i think that is absolutely one of the steps in that direction i mean my sort of challenge to to you because i love playing africa is i agree it'd be great if we could codify all of the constraints around an agent prior to release but do you not just think the demand from xcos and and and business leaders and actually not just demand but actually some of them will just be empowered to create and release agents before the CISO can get their hands on them.
Ash HuntHOST
Do you not just think that it's going to be fixed in production, not prior, because otherwise we're just going to have a backlog of pilots? And I know there's something like 80, 90% of AI pilots don't even reach production anyway.
Ash HuntHOST
And I think that's because the design requirements that we're talking about, right, are halting them in dev and UAT and pre-prod.
Adam RobinsonGUEST
So Elevar is one example.
Adam RobinsonGUEST
They were doing basically ETL, you know, like, uh...
Adam RobinsonGUEST
I mean, they still are.
Adam RobinsonGUEST
They, he sold that business.
Sami HeroGUEST
And, uh, and then I did a whole bunch of work.
Sami HeroGUEST
I was in an ETL startup, and, uh, I actually spent almost twenty-five years traveling the world, and then I came back to Finland, uh, eight years ago.
Joe ReisHOST
Okay, cool.
Joe ReisHOST
Welcome back.
78 more episodes mention Extract, transform, load.
Create an account to see the whole feed, search across every transcript, and follow the entities you care about.