Adding Public Holidays To A Date Dimension

Reza Rad continues his series on Power BI date dimensions:

To get public holidays live, you first need an API that is giving you up-to-date information. There are some web pages that has the list of public holidays. I have already explained in another blog post how to use a web page and query public holidays from there. That method uses custom functions as well, here you can read about that.

The method of reading data from a web page has an issue already; Web.Page function from Power Query is used to pull data from that page, and this function needs a gateway configuration to work. There is another function Xml.Document that can work even without the gateway. So because of this reason, we’ll use Xml.Document and get data from an API that provides the result set as XML. is a great free website with calendars for 36 countries which I do recommend for this example. This website, provides the calendars through XML format. There are other websites that give you the calendar details through a paid subscription. However, this website is a great free one which can be used for this example. is created by User Point Inc.

This was an interesting approach to the problem, one I did not expect when first reading the article.  I figured it’d be some sort of date calculation script.

Fiscal Year Columns In A Power BI Date Dimension

Reza Rad takes a date dimension in Power BI and adds fiscal year details:

As you can see in the image above; June 2017 considered as fiscal year 2017. However, July 2017 is part of fiscal year 2018. So the simple logic can be like this:

if (calendar month >= fiscal year start)

then fiscal year = calendar year

else fiscal year = calendar year + 1

This code is pseudo code. don’t write that exactly in M! Let’s now implement it in M;

If you have to deal with multiple fiscal years (e.g., state and federal government fiscal years), the process is the same, only repeated.

Azure SQL Data Warehouse Patterns

Murshed Zaman shows us a couple of patterns and anti-patterns for Azure SQL Data Warehouse:

Azure SQL DW is a Massively Parallel Processing (MPP) data warehousing service. It is a service because Microsoft maintains the infrastructure and software patching to make sure it’s always on up to date hardware and software on Azure. The service makes it easy for a customer to start loading their tables on day one and start running queries quickly and allows scaling of compute nodes when needed.

In an MPP database, table data is distributed among many servers (known as compute or slave nodes), and in many MPP systems shared-nothing storage subsystems are attached to those servers. Queries come through a head (or master) node where the location metadata for all the tables/data blocks resides. This head node knows how to deconstruct the query into smaller queries, introduce various data movement operations as needed, and pass smaller queries on to the compute nodes for parallel execution. Data movement is needed to align the data by the join keys from the original query. The topic of data movement in an MPP system is a whole another blog topic by itself, that we will tackle in a different blog. Besides Azure SQL DW, some other examples of a MPP data warehouses are Hadoop (Hive and Spark), Teradata, Amazon RedShift, Vertica, etc.

The opposite of MPP is SMP (Symmetric Multiprocessing) which basically means the traditional one server systems. Until the invention of MPP we had SMP systems. In database world the examples are traditional SQL Server, Oracle, MySQL etc. These SMP databases can also be used for both OLTP and OLAP purposes.

Murshed spends the majority of this blog post covering things you should not do, which is probably for the best.

Basics Of Azure SQL Data Warehouse

Minette Steynberg has an article introducing Azure SQL Data Warehouse:

Azure SQL DW is best used for analytical workloads that makes use of large volumes of data and needs to consolidate disparate data into a single location.

Azure SQL DW has been specifically designed to deal with very large volumes of data. In fact, if there is too little data it may perform poorly because the data is distributed. You can imagine that if you had only 10 rows per distribution, the cost of consolidating the data will be way more than the benefit gained by distributing it.

SQL DW is a good place to consolidate disparate data, transform, shape and aggregate it, and then perform analysis on it. It is ideal for running burst workloads, such as month end financial reporting etc.

Azure SQL DW should not be used when small row by row updates are expected as in OLTP workloads. It should only be used for large scale batch operations.

Azure SQL Data Warehouse is fantastic when you’ve got a setup like above and are willing to pay a premium to make things faster.  And with appropriately distributed data, it certainly does get faster.

Managing Hive Slowly Changing Dimensions

Carter Shanklin shows how to manage Type 1, 2, and 3 slowly changing dimensions in Hive:

The most common SCD update strategies are:

  • Type 1: Overwrite old data with new data. The advantage of this approach is that it is extremely simple, and is used any time you want an easy to synchronize reporting systems with operational systems. The disadvantage is you lose history any time you do an update.

  • Type 2: Add new rows with version history. The advantage of this approach is that it allows you to track full history. The disadvantage is that your dimension tables grow without limit and may become very large. When you use Type 2 SCD you will also usually need to create additional reporting views to simplify the process of seeing only the latest dimension values.

  • Type 3: Add new rows and manage limited version history. The advantage of Type 3 is that you get some version history, but the dimension tables remain at the same size as the source system. You also won’t need to create additional reporting views. The disadvantage is you get limited version history, usually only covering the most recent 2 or 3 changes.

The Hive solution is getting closer and closer to a traditional relational warehouse solution.  And on the whole, that’s a good thing.

The Data Lake From 10,000 Feet

Pradeep Menon has a high-level explanation of what a data lake is and how it differs from traditional data warehouses:

With the changes in the data paradigm, a new architectural pattern has emerged. It’s called as the Data Lake Architecture. Like the water in the lake, data in a data lake is in the purest possible form. Like the lake, it caters to need to different people, those who want to fish or those who want to take a boat ride or those who want to get drinking water from it, a data lake architecture caters to multiple personas. It provides data scientists an avenue to explore data and create a hypothesis. It provides an avenue for business users to explore data. It provides an avenue for data analysts to analyze data and find patterns. It provides an avenue for reporting analysts to create reports and present to stakeholders.

The way I compare a data lake to a data warehouse or a mart is like this:

Data Lake stores data in the purest form caters to multiple stakeholders and can also be used to package data in a form that can be consumed by end-users. On the other hand, Data Warehouse is already distilled and packaged for defined purposes.

One way of thinking about this is that data warehouses are great for solving known business questions:  generating 10K reports or other regulatory compliance reporting, building the end-of-month data, and viewing standard KPIs.  By contrast, the data lake is (among other things) for spelunking, trying to answer those one-off questions people seem to have but which the warehouse never seems to have quite the right set of information.

Data Warehouse Automation

Koos van Strien provides some thoughts on data warehouse automation tools:

Currently, I think there are two main approaches to Data Warehouse Automation

  1. Data Warehouse Generation: You provide sources, mappings, datatype mappings etc.. The tool generates code (or artifacts).
  2. Data Warehouse Automation (DWA): The tool not only generates code / artifacts, but also manages the existing Data Warehouse, by offering continuous insight in data flows, actual lineage, row numbers, etc..

The difference might seem small, but IMHO is visible most clearly whenever changes occur in the Data Warehouse – the second class of tools can handle those changes (while preserving history). With the first class of tools provide you with the new structures, but you need to handle the preservation of history yourself (as you would’ve without DWA).

Read on for a contrast of these two approaches.

Dimensional Modeling

Jen Underwood explains the basics of dimensional modeling:

A dimensional model is also commonly called a star schema. It provides a way to improve report query performance without affecting data integrity. This type of model is popular in data warehousing because it can provide better query performance than transactional, normalized, OLTP data models. It also allows for data history to be stored accurately over time for reporting. Another reason why dimensional models are created…they are easier for non-technical users to navigate. Creating reports by joining many OLTP database tables together becomes overwhelming quickly.

Dimensional models contain facts surrounded by descriptive data called dimensions. Facts contains numerical values of what you measure such as sales or user counts that are additive, or semi-additive in nature. Fact tables also contain the keys/links to associated dimension tables. Compared to most dimension tables, fact tables typically have a large number of rows.

Jen’s post was built off of an early SQL Saturday presentation.  It’s still quite relevant today.

Polybase Design Patterns On Azure SQL DW

Simon Whiteley continues his Polybase on Azure SQL Data Warehouse series.  First, he covers data loading patterns:

That’s enough about data loading for now, there’s another major use case for Polybase that we haven’t yet discussed. Many data processing solutions have a huge, unwieldy overnight batch job that performs aggregates, lookups, analytics and various other calculations.

However, it is often the case that this is not timely enough for many business requirements. This is where Polybase can help.

If we have an External Table over the newest files, this will read these new records at query time. We can write a view that combines the External Table with our batch-produced table. This will obviously go a little slower than usual, given it has to read data from flat files each time, however the results returned will be up to date.

Simon then covers the Create Table As Select statement:

In order to utilise SQLDW effectively, we write SQL for our transformations, rather than relying on external tools such as SSIS. This ensures the work is being done by our compute nodes and, therefore, can be scaled up to increase performance.

General best practice, therefore, would be write stored procedures for each of the data movements we want to occur. This allows us to add in auditing, logging etc. But what we’re interested in here is the core data movement itself.

Writing a traditional INSERT statement isn’t the fastest way to get data into a table. There is a special syntax which creates a new table and inserts into it, that is automatically configured for optimal bulk loading, this is the CTAS, or “Create Table as Select” statement.

This is a pair of interesting posts from Simon.

Star-Schema Benchmark With Hive + Druid

Carter Shanklin and Slim Bouguerra run a Hadoop OLAP system running Hive and Druid against the Star-Schema Benchmark battery of queries:

How did we arrive at the query used to build the OLAP index? There is a systematic procedure:

  1. The union of all dimensions used by the SSB queries is included in the index.
  2. The union of all measures is included in the index. Notice that we pre-compute some products in the index.
  3. Druid requires a timestamp, so the date of the transaction is used as the timestamp.

You can see that building the index requires knowledge of the query patterns. Either an expert in the query patterns architects the index, or a tool is needed to analyze queries or to dynamically build indexes on the fly. A lot of time can be spent in this architecture phase, gathering requirements, designing measures and so on, because changing your mind after-the-fact can be very difficult.

One thing I don’t like so much is that they removed the ORDER BY clauses from some of the queries, as making this change makes it more difficult to use these results for “it’s totally not a comparison so don’t sue us Oracle” purposes.


October 2017
« Sep