Blueprism: how to use the replace function in a calculation stage?

Question

I am reading a text from an application using BluePrism. The text has the following structure (the number varies from case to case): "Please take note of your order reference: 525". I need to be able to extract the number from the text. Looking at the calculation stage, there is a replace function: replace(text, pattern, new-text). I want to use this function to replace all alphabetic characters in my text with an empty string to return only whatever is numeric. How can I input that in the pattern? So I want something like this:

Replace([Order confirmation text ], /^[A-z]+$/, " ")

Also, I tried to look for a proper documentation for the VBOs that are shipped with blueprism, but couldn't find any. Does anyone know where we can get documentations for blueprism functions?

Solution

The Replace() function in calculate stage is the simplest possible one. It's not a regex one!

So, if the stirng is always in that format, then you can use:

Replace([Text],"Please take note of your order reference:","")

If the text is not always that standard, then you should rather use a regular expressions. To do that, you need to use an object, that will invoke a regex code.

In the standard blueprism objects, you can find:

Object: Utility - Strings C#
Action: Extract Regex Values

I think there is no Regex Replace action, by default, so if you'd like to, then you have to implement it. Below you can find a code that I am using:

Dim R as New Regex(Regex_Pattern, RegexOptions.SingleLine)
Dim M as Match = R.Match(Text)
replacement_result = R.Replace(Text,Regex_Pattern,replacement_string)

Comments

[AntsaR]

I have also tried using the Utility-Strings VBO, using the Extract Regex action with a regex \d+, but I don't get anything in my output collection.

[AntsaR]

You are right, I was adding white space instead of replacing with empty string, and yes, I only want the numbers. so right now, i have something like this: Replace([Order confirmation text ], [A-Za-z :]+, ""), and when I evaluate the expression, I get the error: Missing data: [A-Za-z :]+. So I am thinking maybe the pattern here needs to be an actual string instead of a regex.

[CAustin]

You still need to put the forward slashes around it

[AntsaR]

Yeah, even with the forward slashes I get the same error. I think the pattern argument here expects a string input.

[AntsaR]

Yep, this is also what I ended up doing. I just specified the exact text as the "pattern", and will have to use the strings utility for better solutions. thanks!

[Celso Lívero]

the number varies from case to case in other words if the number is 54 or 54321 your answer is incorrect

[carusyte]

But the "Extract Regex Values" action only extracts the matched strings instead of replacing that matched string with targeted value. How to accomplish the replacement utilizing regex is still a question.

[Andrzej Kaczor]

Hey, I've added regex relace code to my answer. Enjoy!

[Leni]

@AndrzejKaczor Thanks I found this quite useful. But R.Replace(Text,Regex_Pattern,replacement_string) gives compiler warning. (Access of shared member, constant member, enum member or nested type through an instance) Better user Regex.Replace(Text,Regex_Pattern,replacement_string)