Search specific text on a PDF and get the page number it exists

(1) Share

Report

Posted on by Tomo-O

Hi, I'm a beginner user of Power Automate.

I have around 10 page PDF (ex. page 1 to 10), and I would like to get the page number where the specific text (ex. SS-4) exists, and extract the one page PDF. The problem is the existing page is different for each PDF. Thus, I'm trying to use AI Builder, Power Automate, and Encodian to solve this.

First, I use AI Builder's "Recognise text in an image or a PDF document" and get the all context. Next, I add "Compose" and type the expression which orders search the specific text "ex. A" and get the page number. However I'm not sure what kind of expression I should use.

The text A is the unique text and should be only 1 page out of 10.

Appreciate if you can help me Thanks!

Memo, Because I use Mac, I can't access to Power Automate Desktop.

Categories:

AI Builder

I have the same question (0)

All responses (6)

Answers (0)

antoinec on at

Like (1)

Report

Hi Tomo,

Here is one way you could do it. After extracting PDF contents with the "Recognize text in an image or a PDF document" action, iterate on "results" and "lines" to find a "text" item which matches your criteria. If it matches, execute the relevant action within that condition block.

This is what it looks like in the Flow editor:

And here's an example run with a match:

Is this what you were looking for? (there might be ways to do this without the double "Apply to each" blocks ... if anyone has the info I'm keen to learn :))

Antoine

Was this reply helpful? Yes No
joshfleisher 2 on at

Like (1)

Report

Hello!

To add to this, and I know this was some time ago, but I was interested in being able to pull the whole page that contains a certain word, and creating its own file. Like the example above, I may have 10 pages, but 2 of the PDF pages have the word that means I need the whole document. I need the 2 pages extracted, and its own file created to store. Is there a way to extract the whole page(s)?

Thank you in advance!

Josh Fleisher

Was this reply helpful? Yes No
JRosenboom 2 on at

Like (1)

Report

Hello!
Just on this, in the case of multiple keywords appearing on the same or different pages in a document, is it possible to extract which page each keyword appears on even with duplicates?
e.g.
Keyword appeared on: p.g. 1, p.g. 1, p.g. 2, p.g. 3, p.g. 4, p.g. 4
?

Was this reply helpful? Yes No
theresia 10 on at

Like (2)

Report

hi can u make the dekstop flow version?

Was this reply helpful? Yes No
Lakshya2025 4 on at

Like (1)

Report

Hello,
When i use Split pdf by text it gives me the split pdf not singular pages with the text, is there any way to get the single pages with the specific text so that i can use ai model on those pages.

Was this reply helpful? Yes No
Suggested answer

BoredFish 26 on at

Like (1)

Report
@theresia

PA Desktop does not natively have a function to iterate through pages of a PDF file to extract individual pages where specific keywords are found. However, it does have enough functionality to create your own method of accomplishing this task.

To do this, we first identify the total number of pages in the PDF; then create a fixed quantity loop that looks through each page for specific text; and finally extracting each page that text was found on. Sounds simple until we realize there's also no function to tell us how many pages a PDF has. However, you can obtain the page count by asking it to extract more pages than could possibly exist and thus forcing the flow to create an error. We can then capture the error, extract it's text and use it to identify what page number the function got up to before the error occurred. Here's what that might look like:

**REGION Obtain Total Quantity of Pages in PDF SET MyPDF_FilePath TO $'''YourFilePathGoesHere''' Pdf.ExtractPages PDFFile: MyPDF_FilePath PageSelection: $'''1-1000''' ExtractedPDFPath: MyPDF_FilePath IfFileExists: Pdf.IfFileExists.AddSequentialSuffix ExtractedPDFFile=> ExtractedPDF ON ERROR END ERROR => LastError SET LastErrTrimmed TO LastError.Message.Trimmed.Trimmed Text.SplitText.SplitWithDelimiter Text: LastErrTrimmed CustomDelimiter: $''',''' IsRegEx: False Result=> LastErrSplit SET MyPDF_PageCount TO LastErrSplit[1] Text.ToNumber Text: MyPDF_PageCount Number=> MyPDF_PageCount Variables.DecreaseVariable Value: MyPDF_PageCount DecrementValue: 2 **ENDREGION **REGION Extract Pages LOOP LoopIndex FROM 1 TO MyPDF_PageCount STEP 1 Pdf.ExtractTextFromPDF.ExtractTextFromPage PDFFile: MyPDF_FilePath PageNumber: LoopIndex DetectLayout: False ExtractedText=> ExtractedPDFText Text.ParseText.ParseForFirstOccurrence Text: ExtractedPDFText TextToFind: $'''TextToSearchForGoesHere''' StartingPosition: 0 IgnoreCase: False OccurrencePosition=> Position IF Position > 0 THEN Pdf.ExtractPages PDFFile: MyPDF_FilePath PageSelection: LoopIndex ExtractedPDFPath: MyPDF_FilePath IfFileExists: Pdf.IfFileExists.AddSequentialSuffix ExtractedPDFFile=> ExtractedPDF END END **ENDREGION

It should be fairly easy to adapt that code to fit most use cases. Good luck!

Was this reply helpful? Yes No