# Day: July 27, 2018

rquery is based on an appreciation of Codds’ relational algebra. Codd’s relational algebra is a formal algebra that describes the semantics of data transformations and queries. Previous, hierarchical, databases required associations to be represented as functions or maps. Codd relaxed this requirement from functions to relations, allowing tables that represent more powerful associations (allowing, for instance, two-way multimaps).

Codd’s work allows most significant data transformations to be decomposed into sequences made up from a smaller set of fundamental operations:

• select (row selection)
• project (column selection/aggregation)
• Cartesian product (table joins, row binding, and set difference)
• extend (derived columns, keyword was in Tutorial-D).

One of the earliest and still most common implementation of Codd’s algebra is SQL. Formally Codd’s algebra assumes that all rows in a table are unique; SQL further relaxes this restriction to allow multisets.

rquery is another realization of the Codd algebra that implements the above operators, some higher-order operators, and emphasizes a right to left pipe notation. This gives the Spark user an additional way to work effectively.

They include a fairly lengthy example and give a great introduction to the tool.  It’s now officially on my list of stuff to try out.

Okay, not a perfect score but good enough for me – right now, I’m more interested in the explanations of the model’s predictions. For this, we need to run the `lime()` function and give it

• the text input that was used to construct the model
• the trained model
• the preprocessing function
``explainer <- lime(clothing_reviews_train\$text, xgb_model, preprocess = get_matrix)``

With this, we could right away call the interactive explainer Shiny app, where we can type any text we want into the field on the left and see the explanation on the right: words that are underlined green support the classification, red words contradict them.

I hadn’t used LIME for this before, and it looks very interesting.  H/T R-Bloggers

When you want to get latest (actual) state of data in a temporal table, you can query the same way as you query a normal table. If the PERIOD columns are not hidden, their values will appear in a SELECT * query. If you specified PERIOD columns as hidden, their values won’t appear in a SELECT * query. When the PERIOD columns are hidden, reference the PERIOD columns specifically in the SELECT clause to return the values for these columns.

To perform any type of time-based analysis, use the new FOR SYSTEM_TIME clause with four temporal-specific sub-clauses to query data across the current and history tables.

• AS OF <date_time>

• FROM <start_date_time> TO <end_date_time>

• BETWEEN <start_date_time> AND <end_date_time>

• CONTAINED IN (<start_date_time> , <end_date_time>)

• ALL

Do check it out.  The big downside I’ve seen to temporal tables is that I can’t query a temporal table as of a per-row timestamp in another table.

I have a great interest in hierarchy solutions, as I plan to write a book on implementing hierarchies in SQL Server in the next year or so (after I see what happens in SQL Server vNext after 2017). Something I didn’t include in my design testing when I built my presentation on hierarchies (available from a link here) a few years back was the cost to remove nodes from a tree. I learned a big lesson today about the cost of deleting from a hierarchy (the hard way). It is definitely something I will be digging deeper into at some point in a lab environment, comparing multiple methods of deleting from a hierarchy, but I wanted to write down the issue with self-referencing table for future use (and hopefully for your use too.)

We have a table with 100s of millions of rows, and deletes from this table were taking forever. There were 17 foreign key constraints to this table, but each of the references were only showing as 1% of the plan. The physical DELETE from the clustered index operator was ~80%, but in the query, there showed up an operator that looked ominous.

Chris Webb has a couple interesting posts on using Process Monitor, a venerable sysinternals tool, to troubleshoot issues with Power Query performance.  First, Chris looks at how often Power Query opens and closes a file for data processing:

Troubleshooting Power Query performance issues in Power BI and Excel can be difficult because it’s a bit of a black box: there’s nothing in the UI to tell you what’s going on inside the Power Query engine and the diagnostic logs are very difficult to interpret. With relational data source like SQL Server you can use tools like SQL Server Profiler to see the queries that are being run by Power Query, and I blogged recently about using Fiddler to troubleshoot OData performance issues; but what about file-based data sources, which often present the most challenges regarding performance?

Process Monitor, a free tool from Microsoft, allows you to monitor file system activity in real-time and even having spent a limited amount of time using it I can already tell that it can provide a lot of information to help identify performance issues with file-based data sources. Take, for example, the scenario I described in my recent post on improving the performance of merge operations. In that post (which I suggest you read before you carry on) I mentioned that it looked as though the Power Query engine was reading data from one of the source files multiple times and Process Monitor confirms that this indeed the case.

This post is really just a quick follow-on from my post earlier this week on using Process Monitor to troubleshoot Power Query performance issues with file-based data sources, which I suggest you read before carrying on. I realised, after playing around with Process Monitor some more, that the ReadFile operation actually tells you how much data is being read from a file when a Power Query query is running. For example, here’s a sample of some of the ReadFile operations captured while running the unoptimised version of the query I talked about in my last post:

I get the feeling that really learning a few sysinternal tools can be extremely useful, like how learning Wireshark can make you a better DBA.