This post explains how to remove unwanted characters from a string in a Power Automate flow. These can be regular characters or unprintable characters, such as control codes that you need to strip out. I am going to demonstrate two different ways of achieving this
My example string:
<note> <to>[email protected]</to> <from>Michael Knight</from> <heading>Reminder</heading> <body>Don't forget to clean KITT this weekend!!! </body> </note>
The aim is to remove the apostrophes, exclamation marks and the two non-printable characters. The non-printable characters are visible in Notepad++ :
As you can see from the screenshot, the non-printable characters are the SOH (Start of heading) and BEL (Bell).
Method 1 – Loop through an array of invalid Characters
In the first method an array of invalid characters is created. The array is used as the basis for an apply to each loop, to remove each character in turn. Here is a screenshot of the flow:
Brief overview of how the flow works:
- A variable called string is defined which contains the complete string including the invalid characters.
- A compose action called RemovalChars is created, which is an array of characters that you’d like to be removed.
- An apply to each loop is created using the array RemovalChars as it’s basis.
- Inside the loop, a compose action called “removeCharacter” uses the replace function to remove the character in the current iteration of the loop.
replace(variables('string'), item(), '') - The variable string is updated with the output from the compose action.
- Inside the loop, a compose action called “removeCharacter” uses the replace function to remove the character in the current iteration of the loop.
- Finally, the content of the string with the characters removed is placed inside a compose action, so that you can check if the output is as expected.
In the above example I have only put four characters in the array RemovalChars, like this:
[ "'", "!", "\u0001", "\u0007" ]
The first two elements of the array are simple, but the third and fourth require an explanation. It is not possible to type non-printable characters directly into the flow editor, so you must use the Unicode values that represent them.
- u0001 is the SOH character.
- u0007 is the BEL character.
A note of caution, if you create a long array of unwanted characters like this:
[ "\u0001", "\u0007", "'", "!", "@", "£", "^", "=", ",", "%", "[", "]", ";", "~", "?", "<", ">", "|" ]
The apply to each loop will take a long time to execute. So keep the number of replacements down to the minimum. If you need better performance, you can replace every character in a single action with a nested replace, as explained below.
Method 2 – Use a nested replace action to remove the unwanted characters
Although the method above is perfectly valid, if you have many replacements to do, the performance might not be as fast as you would like. It is possible to execute the same process as the above, but in a single expression without any apply to each loops. Here is an example expression:
replace
(
replace
(
replace
(
replace(variables('string'), '''', ''),
'!',
''
),
decodeURIComponent('%01'),
''
),
decodeURIComponent('%07'),
''
)
A nested replace action can perform the whole process in a single step. Notice that the non-printable characters are now selected with the use of the decodeURIcomponent expression. Here is a screenshot of an example flow doing the same operation in a single step:
Hope this helps, do let me know in the comments how you get on.
Adam W says
I would like to keep the list of special characters in a SharePoint list so that users can easily update them and then pull the special characters from the list into the flow. How can this be done?
Tristan Holt says
In the “Replace Unwanted Character” compose action, what is ‘NewFileContent’?
Mike M. says
Thank you, Paul. Method 2 is exactly what I needed!