Category: ETL / ELT

Cathrine Wilhelmsen continues a series on Azure Data Factory. Since we left off, Cathrine has three new posts. First, a look at how we monitor Azure Data Factory pipelines:

In the previous post, we looked at the three different trigger types, as well as how to trigger pipelines on-demand. In this post, we will look at what happens after that. How does monitoring work in Azure Data Factory?

Now, if we want to look at monitoring, we probably need something to monitor first. I mean, I could show you a blank dashboard, but I kind of already did that, and that wasn’t really interesting at all 🤔 So! In the previous post, I created a schedule trigger that runs hourly, added it to my orchestration pipeline, and published it.

Second, using annotations to customize views:

Annotations are additional, informative tags that you can add to specific factory resources: pipelines, datasets, linked services, and triggers. By adding annotations, you can easily filter and search for specific factory resources.

You need to figure out what kind of annotations make sense to you. Maybe you want to filter on the different logical steps of your solution, so you add the tags extract and transform? Perhaps ingest and prepare? Or maybe you want to tag resources with a business unit or a project name? It’s entirely up to you. All I recommend is that you’re consistent 🙂

That’s a problem for me—the only thing I’m consistent about is inconsistency. Third, Cathrine introduces the different runtimes available to us:

An integration runtime (IR) specifies the compute infrastructure an activity runs on or gets dispatched from. It has access to resources in either public networks, or in public and private networks.

Or, in Cathrine-speak, using less precise words: An integration runtime specifies what kind of hardware is used to execute activities, where this hardware is physically located, who owns and maintains the hardware, and which data stores and services the hardware can connect to.

There’s a lot of good material in each of these three posts.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory by looking at triggers:

One important thing to note is that all times are in UTC. And since UTC does not observe daylight saving time… Well, let’s just say that if you need to execute pipelines during the workday and you have business users waiting for data, you may want to plan some trigger maintenance on the days when you fall back or spring forward. I know. Ugh 🙂 I’m hoping for better timezone support in the future 🤞🏻

Schedule triggers and pipelines have a many-to-many relationship. That means that one schedule trigger can execute many pipelines, and one pipeline can be executed by many schedule triggers.

Time-based triggers aren’t the only options, however—Cathrine also looks at the other three possibilities.

Comments closed

Cathrine Wilhelmsen shows us how to debug Azure Data Factory pipelines:

You debug a pipeline by clicking the debug button:

Tadaaa! Blog post done? 😀

I joke, I joke, I joke. Debugging pipelines is a one-click operation, but there are a few more things to be aware of. In the rest of this post, we will look at what happens when you debug a pipeline, how to see the debugging output, and how to set breakpoints.

Turns out there’s more to it than clicking a button.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory:

The other way to build this solution is by creating an orchestration pipeline with two execute pipeline activities. This gives us a little more flexibility than having a single pipeline, because we can execute each pipeline separately if we want to.

Let’s start by creating a new pipeline and adding two execute pipeline activities to it. In the activity settings, select the pipelines to execute, and check wait on completion:

Read on for the demonstration.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory:

So far in this Azure Data Factory series, we have looked at copying data. We have created pipelines, copy data activities, datasets, and linked services. In this post, we will peek at the second part of the data integration story: using data flows for transforming data.

But first, I need to make a confession. And it’s slightly embarrassing…

I don’t use data flows enough to keep up with all the changes and new features

To be fair to Cathrine, this is a rapidly-changing part of ADF.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory. Catching up from the last time around, we first see the Copy Data activity:

You can copy data to and from more than 80 Software-as-a-Service (SaaS) applications (such as Dynamics 365 and Salesforce), on-premises data stores (such as SQL Server and Oracle), and cloud data stores (such as Azure SQL Database and Amazon S3). During copying, you can define and map columns implicitly or explicitly, convert file formats, and even zip and unzip files – all in one task.

Yeah. It’s powerful 🙂 But how does it really work?

Then Cathrine hits datasets:

But… please, please, please don’t use “source” or “destination” or “sink” or “input” or “output” or anything like that in your dataset names. It makes sense when you have one pipeline with one copy data activity, but as soon as you start building out your solution, it can get messy. Because what if you realize you want to use the original destination dataset as a source dataset in another copy data activity? Yeah… 🙂

So! Let’s rename the datasets.

After that, it’s on to linked services:

Azure Key Vault is a service for storing and managing secrets (like connection strings, passwords, and keys) in one central location. By storing secrets in Azure Key Vault, you don’t have to expose any connection details inside Azure Data Factory. You can connect to “the application database” without directly seeing the server, database name, or credentials used.

Cathrine is rolling with this series and it’s been great so far.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory with a discussion of pipelines:

Pipelines are sorted by name, so I recommend that you decide on a naming convention early in your project. And yeah, I keep saying this to everyone else, but then I can never decide on how to name my own pipelines, haha 🙂 Don’t worry if you end up renaming your pipelines several times while you work on your project. It happens, and that’s completely fine, but try to stick to some kind of naming convention throughout your project.

In addition to naming conventions, you can create folders to organize your pipelines. Click the actions ellipsis next to the pipelines group, then click new folder.

Read on for more.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory. First, we get an overview of the available components:

Pipelines are the things you execute or run in Azure Data Factory, similar to packages in SQL Server Integration Services (SSIS). This is where you define your workflow: what you want to do and in which order. For example, a pipeline can first copy data from an on-premises data center to Azure Data Lake Storage, and then transform the data from Azure Data Lake Storage into Azure Synapse Analytics (previously Azure SQL Data Warehouse).

Then, Cathrine looks at the Copy Data wizard:

LEGO! Yay! I love LEGO. Rebrickable is an online service that will show you which LEGO sets you can build from the sets and parts you already own. Fun! 🙂

They also have a database of all official LEGO sets and parts (including themes and colors) that you can download for free as CSV files or JSON files.

The CSV files are automatically generated at the start of each month and can be found on rebrickable.com/downloads. 

Cathrine takes this LEGO data and feeds it into Azure Data Lake Storage.

Comments closed

Ginger Daniel continues a series on moving data incrementally from SQL Server to Azure Blob Storage:

In Part 1 of this series, we demonstrated how to copy a full SQL database table from a SQL Server database into an Azure Blob Storage account as a csv file.  My client needed data moved from their on premise SQL Server database to Azure, and then needed the daily incremental data changes uploaded as well.  This article will discuss how to upload the incremental data changes to Azure after the initial data load.

Click through for the process.

Comments closed

Cathrine Wilhelmsen continues a series on Azure Data Factory:

In the introduction to Azure Data Factory, we learned a little bit about the history of Azure Data Factory and what you can use it for. In this post, we will be creating an Azure Data Factory and getting familiar with the user interface.

Spoiler alert! Creating an Azure Data Factory is a fairly quick click-click-click process, and you’re done. But! Before you can do that, you need an Azure Subscription, and the right permissions on that subscription. Let’s get that sorted out first.

This post is all about setup and getting an overview of the ADF canvas.

Comments closed