I didn’t actually attend the summit, but rather took advantage of the fact that Microsoft is generous enough to make all of the sessions available on-demand. The Advanced Expressions for MS Flow session, led by Stephen Siciliano, was chock-full of information on the various types of advanced expressions available in MS Flow, suggested uses for these different types, and real-world demos. Using advanced expressions can seriously kick up your MS Flow game, so I highly recommend becoming familiar with these gems.
Advanced expressions are function expressions that can be written in a compose action or inline in a conditional. When an advanced expression is written in a compose action, the result of the expression can then be used in other areas of your flow. On the other hand, if the expression is written in a conditional, the result of the expression will be used to determine to which path the flow will continue. In this case, the expression must begin with the @ character, and the expressions that can be used here include equals(), and(), or(), not(), less(), lessOrEquals(), greater(), and greaterOrEquals().
However, I may be getting ahead of myself. You may still be asking “Where do I find advanced expressions?” So, let me start there. Once you create an action, click on “Add dynamic content”, in the pop up, there are tabs for “Dynamic content” and “Expression”. Click on “Expression”. All of the advanced expressions are listed by type. They also include a short snippet description to assist you in determining what they do and how you might use them. It helps that most of their names are fairly descriptive as well.
The beauty of advanced expressions is they give you many more options for working with your data; and therefore, you can create flows that can do more for you. With advanced expressions you have the power to manipulate and generate data in the following ways:
Converting Data Types
- MS Flow automatically converts some data types. For example, integers will automatically be converted to strings. What this means is that the data will be available both as an integer and as a string in the data content.
- Other, less obvious, conversions can be written as an expression by the flow author. For example, if you need to convert a base64 encoded string to a string, you would use base64ToString(). There are advanced expressions for converting to string, floating point, integer, boolean, base64, Data URI, URI component, binary, array, JSON object, and XML content.
Working with Strings
- There are expressions for joining two strings, concat(); extracting a chunk of a string, substring(); or replacing sections of a string, replace().
- If you need to convert the case of a string, you can use toLower() and toUpper().
- Finding the location of text in a string is easy with indexOf() and lastIndexOf().
- Lastly, if you need to check if a string starts with or ends with a value, you can use startsWith() or endsWith().
- Doing simple arithmetic, or even advanced arithmetic, can be achieved with the arithmetic expressions. These include:
- add(), returns the result of adding two numbers
- sub(), returns the result of subtracting two numbers
- mul(), returns the result of multiplying two numbers
- div(), returns the result of dividing two numbers
- mod(), returns the remainder after dividing two numbers
- Advanced arithmetic is possible, because you can use expressions within expressions. For example, if you want to complete a + b / c, your expression would be add(a, div(b,c)).
- Min() and max() are expressions that return the minimum or maximum value from an array of numbers.
- There are a bunch of advanced expressions for manipulating date/time. These include, but are not limited to, expressions for:
- Returning the current timestamp as a string, utcNow()
- Adding times together, addMinutes(), addHours(), addDays(), etc
- Converting time zones, convertTimeZone(), convertToUtc(), convertFromUtc()
- Formatting date/time into a string, formatDateTime()
- Returning a segment of time from a timestamp, dayOfWeek(), dayOfMonth(), etc.
- The two most helpful examples of expressions that generate data are:
- guid(), returns a globally unique string
- rand(), which returns a random integer within the specified range
The if() expression can be helpful if you are simply trying to add styling or set a value based on a condition. You can write a simple if() expression rather than add an additional layer of conditions into the flow. Multiple layers of nested conditions can become confusing and can get short circuited if a missing property gets referenced.
Properties that are not set or return as null will cause your flow to fail. To avoid this, use the coalesce() function in conjunction with the ? character. The idea is to set a default value for the property that will be used if the property returns as null. This looks something like this: coalesce(body(‘Get_record’)?[‘content’], ‘Default Value’).
There are also some really cool advanced expressions for working with lists/arrays, and workflows. I encourage you to poke around in the Expression tab to see what else is available. In the session, Siciliano also briefly discussed asynchronous actions (using 2 triggers in the same flow), working with other flows (using the Flow Management Connector), and calling nested workflows (use this to break up larger flows into smaller chunks). As I stated earlier, this session was chock-full of information. It was well worth the time spent. I’d like to think my Flow game has been kicked up a notch as a result.