How to select a specific text from web page

[awadheshyona]

Hi All,

I am new to PowerAutomate and this is my first flow. I am trying to extract a specific text value from a web page. Right now I am able to extract the complete sentence using 'Extract data from web page' or 'Get details of element on web page', however I am interested in parsing the specific text value.

E.g. Please accept my order with order no 123456 and phone number 8573454567.

I wan to extract 123456 and 8573454567.

Thanks,

Awadhesh

[Highboy]

If you are sure your order and phonenumber are always 6 and 10 digits you can do it like this after you have extracted the string from the web page. (Use your extracted data from the web page, not the teststring😊)

This is just an example, to give you an idea

[awadheshyona]

Thanks @Highboy  This should help. I will give it a try, however can you please help if there is any way to do it  without parsing it(something like PowerAutomate functionality).

Mostly the length will remain same but they can be in any order and also the sentence may change.

E.g. Accept with phone number 8573454567 and order no 123456 

[Highboy]

Using the one below picks the order and phonenumber correct, regardless of the order in the text.

Picks exactly 6 consecutive digits (?<!\d)(\d{6})(?!\d)

Picks exactly 10 consecutive digits (?<!\d)(\d{10})(?!\d)

You can play around with it on https://regex101.com/

[awadheshyona · Accepted solution]

Thanks @Highboy You have been really helpful.  That will work 90% of time, however I have few scenarios where length is not consistent (I have email id as well which I need to extract) e.g. order 1234 and phone number might have dashes e.g . 857-345-4567 . I also have some cases which even have blank value and I have to extract if that information is blank or not.

e.g. Please accept order 12234 with phone no: 

[MichaelAnnis · Accepted solution]

Look at the Parse Text rules; I am sure you can find an expression that will fit all forms of phone numbers with various parenthesis, dashes, dots, etc.

If you can't, then I would use nested If's for which format you can find.  Ex.

• If Parse Text d10 is empty
  • If Parse Text ddd-ddd-dddd is empty
    • If Parse Text (ddd)ddd-dddd is empty
      • Phone number = "blank"
    • End
    • Get phone number
  • End
  • Get phone number
• End
• Get phone number

Same goes for your other items:

• If Parse Text (for order number) is empty
  • Set variable Order Number to "blank"
• End

Hope this helps.

[Pavel_NaNoi]

Tough one if its as dynamic as you say, I'll also just throw in a couple of suggestions, see if any of them stick:

- If its a UI Element use an "Get details of element on webpage" action and select it, this will save the text.

- Extract all the text and use a parse to match for any digit number occurrences, then calculate the respective length of the numbers, I'll assume here that phone number length is always > than order number length and you can use an if statement decision that simply says what is what based on that condition.

Hope this helps.

[Highboy]

May I suggest some kind of data sanitizing?
But you can get a long way with regex, but you need to know what format you can expect.

Where are the data comming from?

[Highboy]

You could use ^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$ to validate the email - but there are lot's of other regex examples for email extraction.

[awadheshyona]

Hi @Highboy @MichaelAnnis @Pavel_NaNoi  Thanks.All of your suggestions were really helpful . Can you please help me what is wrong in below reg expression for email. I am not getting any value in my variable.

/([a-zA-Z0-9._-]+@[a-zA-Z0-9._-]+\.[a-zA-Z0-9_-]+)/

[Highboy]

Just skip the first and last /
This works for me
([a-zA-Z0-9._-]+@[a-zA-Z0-9._-]+\.[a-zA-Z0-9_-]+)

[awadheshyona]

Thanks @Highboy Worked for me too.

[Powerlearnerd]

Hi Highboy,

Could you explain how the flow should look like when you want to write the result to excel?

[Highboy]

What exactly would you like your flow to do?

[Powerlearnerd]

Hi @Highboy ,

I'd like to search for specific words in a webpage and then store these results in Excel linked to an item. 

• Example: When I have on Amazon 30 pages with 10 items (laptops); per individual item page I want to search wether somewhere on the page the word:

• "i 7th 11 gen"
• and/or "64 GB RAM"

is written down.

I then want to store this info in an Excel by extracting the laptop name and then linking wether one of those two words was on its webpage.

Many thanks in advance for your help