, ,

Introduction to Unit Testing with PySpark.

There are few things in life that are worse then cracking open some serious PySpark pipeline code, and then realizing there isn’t a single function written to encapsulate logic … wondering if some change you are about to make will bring down the whole pipeline. When you are new to a codebase you don’t know what you don’t know, you don’t have any backstory and you are usually flying by the seat of your pants in the beginning. When you have no unit tests, usually the only other way to test changes on a Spark pipeline are to run it …. which is sometimes easier said then done in a development environment. The first line of defence should be unit testing the entire PySpark pipeline.

Read More
, , , ,

What makes MLOps so hard? Thoughts for Data Engineers.

It seems like today the problems and challenges of Data Engineering are being solved at a lightning pace. New technologies are coming out all the time that seem to make life a little easier (or harder) while solving age old problems. I feel like Machine Learning Ops (MLOps) is not one of those things. It’s still a hard nut to crack. There have been a smattering of new tools like MLFlow and SeldonCore, as well as the Google Cloud AI Platform and things like AWS Sage Maker, but apparently there is still something missing. Nothing has really gained widespread adoption … and I have some theories why.

Read more
, , , ,

Introduction to Delta Lake on Apache Spark … for Data Engineers

If you’re anything like me when someone says Delta Lake you think DataBricks. But, the mythical Delta Lake is an open source project, available to anyone running Apache Spark. It seems also too good to be true, ACID transactions on the Spark scale? Incredible. This is the future, it has to be. The lines of what is a data warehouse have been starting to blur for a long time, I have a feeling Delta Lake will be the death blow to the traditional DW … or its rebirth??

Read more
, , ,

The Future of Data Engineering, and the Tale Of Two Data Warehouses.

I am going to peer into the crystal-ball, the seeing stone, looking into the murky future of Data Engineering to see what mysteries it holds. I’ve seen a story, a tale of two Data Warehouses, I’ve seen Machine Learning, Streams, Distributed Systems, Storage, the eternal SQL. A lot has changed in the world of Data Engineering in the last few years, but a lot has not changed in the data world as well. Articles about the end of ETL the rise ELT, Hadoop being dead, new data paradigms, no code data flows, managed services, yet very little has actually changed, or it does at a snails pace. Yet, inevitably the store and future of data engineering can be told through the tale of two data warehouses.

Read more
, , ,

Big Data File Showdown – Part 2 – ORC with Python.

In part 1 of the big data file formats we reviewed Parquet vs Avro. It was apparent from the start that the two file formats were built for different things. Avro is clearly a complex row structured file format used in communication and transactions, where schema is king and nested structures are no problem. Parquet on the other hand has risen to the top with the popularity of Spark, is columnar based storage and is well suited to structured and tabular type data. But, lest the annals of inter-webs call us uncouth and forgetful, we must add ORC file format to the list.

Read more
, , , ,

Intro to Spark ML Pipelines for Data Engineers

Don’t you like stuff for free? Don’t you like it when stuff I just handed to you? I mean when is that last time you didn’t want to get a free t-shirt. How about 20 bucks in the mail from you Grandma? That’s kinda what Pipelines are in Spark ML. The Apache Spark ML library is probably one of the easiest ways to get started into Machine Learning. Leaving all the fancy stuff to the Data Scientist is fine, Data Engineers are more interested in the end-to-end. The Pipeline, and the Spark ML API’s provide a straight froward path to building ML Pipelines that lower the bar for entry into ML. So, set right up, come get your free ML Pipeline.

Read more
, , , ,

Converting CSVs to Parquets… with Python and Scala.

With parquet taking over the big data world, as it should, and csv files being that third wheel that just will never go away…. it’s becoming more and more common to see the repetitive task of converting csv files into parquets. There are lots of reasons to do this, compression, fast reads, integrations with tools like Spark, better schema handling, and the list goes on. Today I want to see how many ways I can figure out how to simple convert an existing csv file to a parquet with Python, Scala, and whatever else there is out there. And how simple and fast each option is. Let’s do it!

Read more
, , ,

Top 10 Data Engineering Blogs

I’ve always been surprised with the rise of data engineering and big data, how hard it is to find good data engineering content that is somewhat regular. Tech moves fast and I feel like data engineering moves even faster. There are always new tools and systems coming out with regular frequency, it’s hard to keep up with what’s hot and whats not. But, I still think it’s important to keep a finger on the pulse of what tech stacks are starting to take over (Spark) and what is fading into oblivion. So here is my top ten list of data engineering blogs, these are the places that I frequent so I at least know what’s going on in the world of data engineering.

Read more
, , ,

Musings on Python’s map() and filter()

I’ve always been surprised at the distinct lack of most Python code I’ve seen using the map() and filter() methods as standalone functions. I’ve always found them useful and easy to use, but I don’t often come across them in the wild, I’ve even been asked to remove them from my MR/PR’s, for no other reason then that they are supposedly ambiguous to some people? That’s got me thinking a lot about map() and filter() as related to readability, functional programming, side effects and other never ending debates where no one can even agree on the “correct” definition. Seriously. But, I will leave that rant for another time.

Read more
, , , ,

4 DataWarehouse-ish Functions For Your PySpark Dataframes

Ever felt like just exploring documentation… seeing what you can find? That’s what you do on a cold, first snowstorm of the year Sunday afternoon. After the initial fun has warn off, the kids don’t want to go outside anymore, and Netflix has nothing new to offer up. So I thought I might as well spend some time poking around the PySpark Dataframe API, seeing what strange wonders I can uncover. I did find a few methods that took me back to my SQL Data Warehouse days. Memories of my old school Data Analyst and Business Intelligence days in Data Warehousing… the endless line of SQL queries being written day after day. Anyways lets dive into the 4 analytical methods you can call on your PySpark Dataframe, buried in the documentation like some tarnished gem.

Read more