Dynamic file viewer (SP Webpart) within a flow

[PowerAutoGuy1]

Hey folks,

hey @Expiscornovus :D,

Here is an overview of my task:

I need to copy a bundle of sites for a new workspace at the same site and put those to the Navigator within a Flow.

I managed that already.

Now I need to change the embedded files inside the new pages in the SP file viewer webparts.

In order to do that, I created a flow that gives me an Excel sheet with the ETAGs and the path of the new documents.

Due to further research, I recognized, CanvasContent1 is different in each site and I'm not powerful enough to recognize the difference.

So I've created a flow to face this problem (I guess):

I've tried my best to translate everything.

So if the body of the CanvasContent1 contains the word file viewer, the content has to be adjusted.

Is it possible to just edit a part of the CanvasContent1 or do I have to write everything inside this item?

I hope I have presented my problem clearly and you can see how much of a not professional I am.

Thanks for your eventual help!

Your biggest fan @Expiscornovus !

PowerAutoGuy1

[Expiscornovus]

Hi @PowerAutoGuy1,

I believe it is a good start to check if the page contains a File viewer web part. Your approach looks valid.

However, I would prefer to check it based on the webpart id. I believe you can check the CanvasContent for the webPartId b7dd04e1-19ce-4b24-9132-b60a1c2b910d. That should be the File Viewer web part id.

After that I would retrieve the page and use a bunch of replace functions to replace the old with the new values for fields like (file, listId, siteId, uniqueId & webId of the file used in the File viewer web part).

Btw, I am now using the old school REST API in this example. But this should also be possible with the new List webparts method of the Graph API.

@equals(item()?['webPartId'], 'b7dd04e1-19ce-4b24-9132-b60a1c2b910d')

[PowerAutoGuy1]

Hey @Expiscornovus,

Thanks for the quick reply. I hope you also had a happy new year.

Now it is time to get back to business.

How do you mean "retrieve the page and use a bunch of replace functions"?

Is there a HTTP-Request that is used for that given purpose?

I'm currently trying to get all the information the file viewer needs.

Is it true that file & listId are informations from the linked documents and siteId, uniqueId & webId are informations recived from the Sharepoint sites?

Thank you again for your time!

Your biggest fan,

PowerAutoGuy1

[Expiscornovus]

Hi @PowerAutoGuy1,

Replace is a function in the expression language. You can add those via the expression editor.

Below is an example

In this example I am retrieving the document properties (like uniqueid, listid, webid, siteid) from a table in a source Excel spreadsheet with a Get a row action.

With a GET request I am retrieving the same properties of the current document used in the file viewer web part of the page. When I know those properties I can execute a "find and replace" kind of expression.

The output is used in a POST/MERGE request in a Send an HTTP request action.

The expression I used for the find replace is below.

I used a couple of split(s) and slice(s) functions to find the current values from the GET request. The replace functions are used to replace it with values form the Excel sheet row.

replace(replace(replace(replace(replace(replace(replace(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'file":')[1], '",')[0], 1), outputs('Get_a_row')?['body/file']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'listId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/listId']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'webId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/webId']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'uniqueId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/uniqueId']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'siteId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/siteId']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'serverRelativeUrl":')[1], '",')[0], 1), outputs('Get_a_row')?['body/serverRelativeUrl']), slice(split(split(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'], 'wopiurl":')[1], '"}')[0], 1), outputs('Get_a_row')?['body/wopiurl'])

[PowerAutoGuy1]

I'm going to try this out.

Will report here when my flow is ready.

THANK YOU

[PowerAutoGuy1]

Hey @Expiscornovus, 

I finally managed to do a flow similar to your explanation. I put the whole flow in a loop because there are multiple documents to change. 
I got the same excel table as you have created but the replace function is not changing the file viewer webpart. Is there something to change inside the loop?

Thank you,

PowerAutoGuy1

[Expiscornovus]

Hi @PowerAutoGuy1,

Yes, if you are using a loop that replace expression probably would not work because it is referring to a Get a row action. Within a loop you can probably use the item() function instead. 

It would help if you share a screenshot of your current flow setup. This way I can make a more accurate suggestion.

Based on what I have read about your setup I would suggest to try and change the format of the fields to use that item function.

Below is the original format of the file field for example

outputs('Get_a_row')?['body/file']

Try and change that to:

item()?['file']

[PowerAutoGuy1]

Hey @Expiscornovus 

I tried to display the important part of the flow as good as possible.

1) Filter Site for file viewer

2) Condition

3) Change file viewer

 

The whole flow works as it should until the last step. I can't figure the problem.

Thank you again.

PowerAutoGuy1

[Expiscornovus]

Hi @PowerAutoGuy1,

Does the last step time-out of does it give you a specific error? If so, can you share that error?

I see you kept the Get a row update approach. What did you use in the replace expression. The original approach of referencing to the fields in the outputs of the get a row action?

Can you share a screenshot of the Excel sheet you are using? This way I can try and reproduce your whole flow setup. This can be useful for troubleshooting your issue.

[PowerAutoGuy1]

Hey @Expiscornovus,

1. no the flow is running successfully but it does not change the files inside the file viewer.

2. this is my replace function:

replace(replace(replace(replace(replace(replace(replace(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'file":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/file']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'listId":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/listId']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'webId":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/webId']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'uniqueId":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/uniqueId']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'siteId":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/siteId']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'serverRelativeUrl":')[1], '",')[0], 1), outputs('Zeile_abrufen')?['body/serverRelativeUrl']), slice(split(split(outputs('HTTP-Anforderung_an_SharePoint_senden_-_Get_Page')?['body']['CanvasContent1'], 'wopiurl":')[1], '"}')[0], 1), outputs('Zeile_abrufen')?['body/wopiurl'])

3. The Excel sheet is structured in exactly the same way as the example you have inserted above. The only difference is an additional column at the beginning of the sheet, named "Identifier", where the respective documents are displayed with file extension (e.g. doc1.docx). This is used in "Get a row" as the Key Column and the Key value. The columns are: Identifier, file, siteId, uniqueId, serverRelativeUrl, wopiurl, webId, listId.

4. I suspect that the problem is that different web parts is implemented on the pages and the HTTP request does not understand what exactly it should change. Can this be?

[yxxxyxcyxcycyxc]

Hey @Expiscornovus , @PowerAutoGuy1 ,

This problem is also in my interest!

Hope you guys will work out a solution!

[Expiscornovus]

Hi @PowerAutoGuy1,

Would it be possible to get the page CanvasContent1 property value first. So, I can see what currently is on the page for which the flow is failing/doing nothing? This way I can troubleshoot why it does not work. If you can share that value as a code snippet that would be great.

To be clear, in my setup it was only a single file viewer web part. No other web parts. I will try and see if I can use the same setup as in your test and see if I can reproduce the issue and find a solution.

[PowerAutoGuy1]

Hey @Expiscornovus,

I think that this way we tried until now is not gonna be the solution. Do you think writing a PATCH-Request with the new Informations within CanvasContent1's "links" and "properties" will solve the problem?

PowerAutoGuy1

[Expiscornovus]

Hi @PowerAutoGuy1,

Just managed to update a couple of file viewer (or should I say File and Media) web parts. So, it should be possible with the Send an HTTP request I shared earlier 😁

The approach which worked for me.

Identify the number of File and Media web parts in the CanvasContent1 property. Loop through those items and replace the correct properties one by one for each item. Store that replacement content in an temp array and join these items back together as one CanvasContent1 string value.

I also used a counter which is used to identify the correct file in the Excel Sheet, based on the Identifier column

Below is the setup

1. Added a Counter (integer variable), CanvasContent1 (array variable)

2. Filter current page content for File and Media web parts

json(outputs('Send_an_HTTP_request_to_SharePoint_-_Get_Page')?['body']['CanvasContent1'])

@equals(item()?['webPartId'], 'b7dd04e1-19ce-4b24-9132-b60a1c2b910d')

3. Loop through the Body of the Filter Array

4. Get a row via a counter

5. Append to array variable. 

replace(replace(replace(replace(replace(replace(replace(string(item()), slice(split(split(string(item()), 'file":')[1], '",')[0], 1), outputs('Get_a_row')?['body/file']), slice(split(split(string(item()), 'listId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/listId']), slice(split(split(string(item()), 'webId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/webId']), slice(split(split(string(item()), 'uniqueId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/uniqueId']), slice(split(split(string(item()), 'siteId":')[1], '",')[0], 1), outputs('Get_a_row')?['body/siteId']), slice(split(split(string(item()), 'serverRelativeUrl":')[1], '",')[0], 1), outputs('Get_a_row')?['body/serverRelativeUrl']), slice(split(split(string(item()), 'wopiurl":')[1], '"}')[0], 1), outputs('Get_a_row')?['body/wopiurl'])

6. Increment the counter

7. Outside the loop join everything back together

{
"CanvasContent1": "[@{replace(join(variables('CanvasContent1'), ','), '"', '\"')}]"
}

[PowerAutoGuy1]

Hey @Expiscornovus ,

The uri i am calling is _api/web/lists/getbytitle(`Site Pages‘)/items(391)

the Body is the Same as in your example. When I run the flow, it deletes all webparts on the page. Do you know why? Is Post the right request type?

[Expiscornovus]

Hi @PowerAutoGuy1,

Does it delete all web parts, or only the File Viewer web parts? Which other web parts do you have on the page?

Please share you current page setup (how many web parts, which type of web parts, etc. This helps with troubleshooting/reproducing your issue).

Maybe we just need to spot the differences between your flow setup and mine as well 😁. 

You also used the square brackets around the replace expression (the CanvasContent1 value) in the POST request, correct?

And the Get Page Send an HTTP request action in my setup just retrieves the whole page without metadata. And I am using the sitespages/pages method for this. Not the getbytitle/items one.

[PowerAutoGuy1]

Hey @Expiscornovus, 

After I run the flow (successfully), all Webparts are getting deleted. (Status Code: 204)

The first page got a Button, A MS Forms Survey, the file viewer and a slicer

the second page got a File Viewer and a picture

the third page got no file viewer.

My flow is containing two loops one to get the different pages for the "get pages" call and the other one is similar to yours. I tried to put the "Update File Viewer" step inside the second loop.

Another Difference is: Your URI is _api/web/lists/getbytitle('Site Pages')/items(140)/CanvasContent1

My URI is: _api/web/lists/getbytitle('Site Pages')/items(140)

because CanvasContent1 is called in the body. With CanvasContent1 in the Body and in the URI i get some property called "value" and nothing chages.

I also tried exactly your solution and the result is the same.

This deleted just the page contents which had a file viewer implemented, which is a good thing i guess. If i put it outside the second loop it deletes all page content on all sites.

The POST-Request is built like yours.

21

Is there a possibility to "PATCH" the different values inside the CanvasContent FileViewer Part?

We can do this 😀

Greetings!

[Expiscornovus]

Hi @PowerAutoGuy1,

One of the differences is that I am only testing/updating a single page. That example page has an ID 140. 

Each page has its own unique ID value. I was assuming you are retrieving the IDs of your pages dynamically and using that unique ID in the GET request, is that the case? 😃

If you are using the number 140 for all the requests, you are basically overwriting all your other pages with the content of a page in your environment with the ID 140. Which probably would explain the strange behaviour you are experiencing.

[PowerAutoGuy1]

Hey @Expiscornovus,

yes I am retrieving the IDs of my pages dynamically. I have no idea why my page is not accepting the changed CanvasContent1 and deletes all contents.

[Expiscornovus]

Hi @PowerAutoGuy1,

Thanks for confirming that you are retrieving those ids dynamically.

I will try and reproduce your issue on my environment. I will change my setup to use the same web parts and use three pages as well.