Change the delimiter when creating a CSV file

[Anonymous]

I am querying a SharePoint List and using the Create CSV Table action to generate a CSV file.

However I want the file to be pipe("|") separated.

Is there a way to change the delimiter when creating a CSV file. I would greatly appreciate any advise on this.

[v-yamao-msft]

Hi @Anonymous,

By default, a comma is used to separate the items. If you want the file to be pipe separated, you could use a replace function to replace the comma with a pipe.

Under the Create CSV table, add a Compose action, and use the following code:

replace(body('Create_CSV_table'),',','|')

Then in the email body, select the output from the Compose action.

Please have a try with it on your side.

Best regards,

Mabel Mao

[Anonymous]

HI Mabel,

Thank you for the reply but the solution proposed by you will also replace any comma in the input file with a pipe ('|').

I just want the separator to be changed from comma to pipe while keeping the "commas" elsewhere intact.

[nihkel]

Any updates on this topic?

This is 2 years old and a pretty basic feature.

Has anyone come up with a working solution that doesn't reaplace all particular characters in the CSV but only the delimiter/separator?

Thanks in advance for your help,

Best regards

[Anonymous]

I am also looking for a solution for this. Any update?

[nihkel]

I was kindly pointed to some potential workaround in a similar thread:

https://powerusers.microsoft.com/t5/Power-Automate-Ideas/Changing-the-delimiter-in-Create-CSV/idc-p/...

Hope this helps @Anonymous 

[Admin_A · Accepted solution]

If you intend to do this with the data contained in a SharePoint List, you could :
1) Get your elements with the SP connector;

2) Initialize a string variable with the columns names you would like, e.g. "name;surname;email" and by hitting enter to add a carriage return;

3) Use "Apply to each" on the value element returned by the SP Get elements block;

4) In the apply to each loop, use an "Add to string variable", where the content would be the dynamic fields extracted from SP, separated with ";" and with a carriage return at the end of the line (hit enter)

Then you can add your string variable as the content of an attachment for instance.

Hope this helps !

[nihkel · Accepted solution]

Hi @Admin_A 

thank for your reply.

I had taken a similar route in order to reach the same destination.

I'm creating the CSV table as intended by MS Flow, but I'm using 'custom columns'.

On each header/column I'm adding some special characters (note that these are also forbidden in the SP list, via column validation).

Note the pipes encapsulating the content (everywhere except the line jump, since I'm using the default carriage returns on the CSV).

My aim was to create an easily replaceable set of characters (in my case, pipe-comma-pipe) I could look for, and then replace it as a whole for the character I need to be used as separator.

So next step is a compose action with a replace expression for the 'character set' I created.
I look for |,| and then replace it with my desired separator.
I also didn't needed to take the carriage retun into account since I'm using the default ones created for each line on the CSV.

I hope thi can help someone 🙂

[Runner55552]

Thanks, this was very helpful. We did almost the same thing, although for some reason, a single replace statement looking for |,| did not clean all the records. Where we had comment fields with commas internally, we needed two more replace statements.  We also switched to the ^ symbol, since some of our comment fields also contained pipes! So our three replace/compose statements after the create csv table action were: 

replace(body('Create_CSV_table'), '^,^', '^')   This replaced ^,^ with a single ^

replace(outputs('Compose'), ',"^', '')  This searched for "^, and replaced it with nothing

replace(outputs('Compose_2'), '"', '')  This searched for a double quote mark, and replaced it with nothing.

[takolota]

Hello All,

I ran into similar issues when pulling CSV data in from Power Automate Desktop with commas in the actual data. I wanted to change the delimiter so I could more easily parse the data without commas in the actual data messing things up.

Here's a link to the scope/flow I built so anyone can easily select their CSV data, enter a new delimiter, and get the new delimiter separated data from the final compose action without the usual errors. The file data just needs to have double quotes around any of the records with commas in them or be converted to a text file that automatically applies quotes to those records.
https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Change-CSV-File-Data-delimiter/m-p/14429...

[arielbelaus]

Hello all!

I'm exporting a CSV from a paginated Report and I need to replace the comma delimiter to ~. I tried with a compose formula replacing the , with the ~ but I have some data with , and I don't want to replace those.

Any ideas or solutions about it?

Thanks

[takolota]

@arielbelaus 

You can try my template for changing the CSV delimiter here:

https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Change-CSV-Data-Delimiter/m-p/1442954#M5...

It’s set to work even if there are commas in the data. You may have to copy the rest of your flow into the template flow though, as expressions in the template scope are often messed up when you try to copy it into an existing flow.

[arielbelaus]

I could figure it out by changing the commas of the columns for another symbol in my query. Then, with a step of my flow, change the delimiters, and then with other step in the flow change to commas the symbol that I put with the query

Thanks!

[CaserUni]

It is feasible workaround, when create files with a limited number of columns. When creating multiple flow with more than 100 columns, it become very time consuming.

Is it possible to create these flows directly in a Jason format? It is two time-consuming with all these clicks and wait.

[DEVREYES]

Oh my God! This is so brilliant! I cannot isolate the real commas in the text from the commas made by CSV table. And this! This is such a blessing that saves me from recoding my excel script. I am creating a CSV to excel with my pre-defined xlsx template and I need a delimiter that is very unique to recognize when to put a text on the next column. It's a big problem when the sentence has it's own comma that was created by the user, it tends to separate the texts after a comma. It's been a my problem for a year now. Here is my final table. Thanks a lot for this! Cheers to you!

[nihkel]

Hey @DEVREYES ,

I'm glad this post is stil helpful after all this time 🙂

Good luck with your flows,

Cheers!

[KenSkinner]

Thank you very much.  I'm always amazed after hours of going down various complicated rabbit holes only to find a simple solution like yours that works!

[Anonymous]

I attempted this way. I got seperated columns (finally) but there's the extra column with ",". Is there a way to fix this or go around it?

[nihkel]

Hi @Anonymous ,

could you show the csv in plain text? like on notepad++ or something?
I'm not sure where that extra column may be coming from. It seems you may have an extra set of separators somewhere.

Also, please expand the 'create table' and the 'compose' steps to take a better look and try to assist you.

Cheers 🙂 

[m_jeppesen]

@nihkel  This is a good solution, but there is 1 flaw that needs to be handled 🙂

If any cells have "," in it, fx. 3,000.00 that cell will be formatted like "3,000.00" by PowerAutomate.

So the csv file will end up looking like this - notice the CreditLimits with ""

ID,Name,CreditLimit,Status
1,John Doe,"3,000.50",Active
2,Serena Young,"1,000.00",Active
3,Henry Ford,500.00,Inactive

From a pseudo point of view, the Create CSV file action will format the file like this, if following your solution:

ID|,|Name|,|CreditLimit|,|Status

Above will result in a file looking like this - notice the CreditLimit will surround the cell value in " ", which will cause the | to be within the " "

ID|,|Name|,|CreditLimit|,|Status
1|,|John Doe|,"|3,000.50|",|Active
2|,|Serena Young|,"|1,000.00|",|Active
3|,|Henry Ford|,|500.00|,|Inactive

We then apply your solution, which will replace all |,| with ; - will have the file looking like this.

ID;Name;CreditLimit;Status
1;John Doe|,"|3,000.50|",|Active
2;Serena Young|,"|1,000.00|",|Active
3;Henry Ford;500.00;Inactive

That will replace all |,| with ;   -   but it will not handle the scenario where a , is in a cell value.

So what needs to be done, is to replace all 3 cases that's happening (insert csv instead of 'triggerBody')

|,|

replace(triggerBody()['text'],'|,|','')

ID;Name;CreditLimit;Status
1;John Doe|,"|3,000.50|",|Active
2;Serena Young|,"|1,000.00|",|Active
3;Henry Ford;500.00;Inactive

,|

replace(triggerBody()['text'],'|,','')

ID;Name;CreditLimit;Status
1;John Doe|,"|3,000.50|";Active
2;Serena Young|,"|1,000.00|";Active
3;Henry Ford;500.00;Inactive

|,

replace(triggerBody()['text'],',|','')

ID;Name;CreditLimit;Status
1;John Doe;"|3,000.50|";Active
2;Serena Young;"|1,000.00|";Active
3;Henry Ford;500.00;Inactive

That will have the file looking like this:

ID;Name;CreditLimit;Status
1;John Doe;"|3,000.50|";Active
2;Serena Young;"|1,000.00|";Active
3;Henry Ford;500.00;Inactive

So last step would be to remove the remaining pipes | 

replace(triggerBody()['text'],'|','')

...which will have the file looking like this:

ID;Name;CreditLimit;Status
1;John Doe;"3,000.50";Active
2;Serena Young;"1,000.00";Active
3;Henry Ford;500.00;Inactive

All above 4 replaces can be nested into 1

replace(replace(replace(replace(triggerBody()['text'],'|,|',''),'|,',''),',|',''))

Really props to you @nihkel for coming up with the idea, I just put the last puzzle pieces together 👍

Home more people can benefit from this.

I'll be more than happy to explain in deeper details, if alle the expression is too confusing.