Sliting multiple lines

[slyuso]

Intialize a variable named "A" of type string and enter a couple of lines of text. Then use Compose to split it. Run.

Then, check execution.


This was expected.
But the next was not. See result below:

I also tried use:
split(variables('A'),'
')
This was accepted as expression but result was the same.

split(variables('A'),'\r\n')
produced the same result.

What is wrong here ?

P.S. My ultimate goal is to parse a multi-line CSV text file from Sharepoint document library. Yes, I'm aware of Excel connector. 

[v-micsh-msft · Accepted solution]

Hi @slyuso,

Thanks for the feedback.

I think the issue here should be the

split(variables('A'),'
')

As the 

'
'

Will not be considered as a valid character in flow expression, per my testing.

To workaround this, please take a try to create an example first, then get the New Line character, after that, use the NewLine Compose output as the Split string:

Follow the steps below:

1. Create a string variable as below, make sure to enter two spaces after Line1, then press enter:

Line1  
Line2

2. Add Compose, under the Input, enter the expression below:

take(first(skip(split(Variables('A'),'  '),1)),1)

3. After that, we could use the Outputs of the Compose to split the string with the Line break character:

split(variables('A'),outputs('Compose'))

4. The running results:

Regards,

Michael

[slyuso]

Thank you, Michael
Your post inpired me for even a simpler solution.


To define variable CRLF, I pressed ENTER key. This is not obvious from the picture 
And it worked  as in your example.
Thank you for helping.

[Gaolai]

Brilliant!

[Anonymous]

Awesome!

Thank you

[juresti]

This will be very useful in working with csv files.

I'm still working on and searching for a solution to do this same thing with TAB delimited.

Update: I just realized you can press tab in notepad then copy it into flow and that lets you split tab delimited files.

[JoseZald101]

This could also work, but I have not tried it. This is the original post

https://powerusers.microsoft.com/t5/Building-Flows/Replace-Newline-in-Flow-Expression/td-p/57333

Converting it to an array should something like this:

split(uriComponent(body('Html_to_text')), '%0A')

 If someone could confirm, please let reply.

[nutansh07]

Hi Jose,

This works too but as we are encoding the content, we need to again decode while parsing the results. 

I tested this in a recent flow 😊

Thanks,

Nutan

[nsiva]

Hi @JoseZald101 ,

It worked. Thank you very much for pointing this out. However below expression,

split(uriComponent(body('Html_to_text')), '%0A')

 returns an array with uri components within it. I need to use uriComponentToString method to change each array item.

uriComponentToString(split(uriComponent(body('Html_to_text')), '%0A')[0])

[aberry]

This is genius - the CRLF solution.

[gilroittoperm]

This is brilliant, simple and it works!

[martinushamers]

Great, works like a charm indeed

[gdb789]

Just making minor tweaks to the smart solutions from many contributors, it helps to check the raw texts of the file you want to parse via a 'compose'

E.g. 

The splitting character may vary depending on the datasources, such as LF (\n) or CRLF (\r\n); in the example above I want to split at '\r\n', which can be constructed via the decodeUriComponent function:

So for each combination:

\r => decodeUriComponent('%0D')

\n => decodeUriComponent('%0A')

\r\n => decodeUriComponent('%0D%0A')

In my case the target CSV email body is followed by some empty lines and then an email disclaimer; I skip it before CSV parsing using this compose (which is 2x CRLF)

Then get the 1st array element from the split:

And finally I can split the CSV body into multiple lines with:

The final output:

HTH