Filtering Array with Partial Text Value

[NR2023]

I am trying to build a flow that allows the user to upload a PDF and then it splits the PDF based on a site code. To accomplish this I am utilising the Adobe API and this guide. However I can't get it to work for my use case.

The issue I have is that the only field I can filter on begins with a random word e.g. HR/IT/Finance followed by the site code. In my flow I have a 'Filter Array' that filters 'Text' containing '(Site Code)' however the flow won't accept it and spits on this error message.

failed: 'The template language function 'contains' expects its first argument 'collection' to be a dictionary (object), an array or a string. The provided value is of type 'Null'.'.

Does anyone have any ideas or any ideas on how I can split a PDF based on a value contained on the page?

[Nived_Nambiar]

@NR2023 

I have created a demo flow

Here i have hardcoded sitecode in compose action like below

Now for json array purposes, i have intialized array 

Here my aim is to filter the B field with values starting either with IT{Sitecode} or HR{Sitecode} or Finance{Sitecode}

So I have used a filter array with or condition like below

or(startsWith(item()?['B'],concat('IT',outputs('SiteCode'))),or(startsWith(item()?['B'],concat('HR',outputs('SiteCode'))),startsWith(item()?['B'],concat('Finance',outputs('SiteCode')))))

(Here I am checking whether the b field value is starting with IT{Sitecode} or b field value is starting with HR{Sitecode} or b field value is starting with Finance{Sitecode} using OR condition)

Hope this help you 

Mark it as solution if it helps 🙂

[AlexEncodian]

@NR2023 

You may be able to use Encodian's Split PDF by Text action. It allows you to use a regular expression to find the correct text value to split the PDF by and is therefore very flexible. 

An example is shown in this blog post.

[NR2023]

Hi,

Thank you for your response. I've tried to modify the code using the example provided but I get the exact same error.

' failed: 'The template language function 'contains' expects its first argument 'collection' to be a dictionary (object), an array or a string. The provided value is of type 'Null'.'.

[Nived_Nambiar]

Hi @NR2023 

But we are not using contains function anywhere in the expression right ?

Only i have used startswith there ?

Could you check ur expression once again or if possible share the expression which you have written 

[NR2023]

Hi Nived,

Maybe it would help if I show you what flow I'm using. You can download the flow from here 

Here is the Parse JSON bit:

In my flow example, I am using this PDF file - Download here 

Now imagine in my scenario, that this PDF contained all the invoices for the company and it had several sheets per supplier. I want to be able to filter on text so that all invoices by company marked 'Halliby' are exported into a separate pdf file.

For this I want to try and do the following:

However, when I try it this is what I get:

The execution of template action 'searchArray' failed: The evaluation of 'query' action 'where' expression '@contains(item()?['Text'], 'Halliby')' failed: 'The template language function 'contains' expects its first argument 'collection' to be a dictionary (object), an array or a string. The provided value is of type 'Null'.'.

When I inspect the output of the previous action, I can see that the field is exported from the PDF correctly:

So that's from the example scenario. I'm basically trying to do similar except our site codes are in the format of Finance (Hero)

where (Hero) is our site code. 

[VictorIvanidze]

Have a look at this: https://powerusers.microsoft.com/t5/Using-Flows/Extracting-a-text-value-from-an-embedded-string-valu...

[VictorIvanidze]

Hi @NR2023,

 as far as I can understand, you are looking for this:

where:

outputs('Recognize_text_in_an_image_or_a_PDF_document')?['body/responsev2/predictionOutput/results'][0]?['lines']

 and

item()?['text']

[VictorIvanidze]

Hi @NR2023,

any comments? 

[NR2023]

Hi Victor,

Sorry I'm not seeing how I can use that code sample in my flow? I'm not using the 'Recognise Text' function I am using the 'Adobe Extract PDF Content' option.

[VictorIvanidze]

Try to use 'Recognise Text' just to make sure it will do what you want.